Switching control method, system, medium, and product for a powertrain

Abstract

The application discloses a switching control method, system, medium and product of a power system, and the method comprises the following steps: collecting the running state data of an electric-hydrogen hybrid power system and calculating the change amount and change rate, combining the three into a state feature sequence and inputting the switching prediction model, processing the hidden layer of the model to obtain a hidden state feature, the hidden state feature is used to predict the switching occurrence probability, the to-be-switched behavior category and the switching remaining time by each prediction branch of the model, if the three satisfy the switching rule, the to-be-switched control parameters corresponding to the to-be-switched behavior category are searched in the mode parameter mapping table; a switching control strategy is generated based on the switching remaining time, the to-be-switched control parameters and the running state data, and the system is controlled to switch. The application can solve the problem that it is difficult to predict the impact caused by switching and actively inhibit the impact before switching when the electric-hydrogen hybrid power system switches in different behavior modes, so as to improve the stability of the power system in the behavior switching control process.

Description

Technical Field

[0001] This invention relates to the field of energy management technology for power systems, and more particularly to switching control methods, systems, media, and products for power systems. Background Technology

[0002] Hydrogen-electric hybrid power systems are complex power systems that integrate fuel cells and power batteries for coordinated energy supply. Under different task phases, load states, and energy supply conditions, they typically need to switch between multiple behavior modes (such as steady-state power supply, peak power compensation, and regenerative braking) to optimize system efficiency, response speed, or energy utilization strategies. However, the switching between different behavior modes is often accompanied by abrupt changes in power reference values, energy distribution ratios, actuator control laws, or energy transfer paths, easily leading to transient energy surges. These energy surges can not only trigger overcurrent and undervoltage protection actions, causing unintended shutdowns or derating, but also exacerbate thermal stress and mechanical fatigue in power devices and transmission components, reducing system reliability and lifespan.

[0003] Existing technologies employ smoothing control methods based on filtering and slope constraints to control the switching of electric-hydrogen hybrid power systems. However, these methods have significant drawbacks: their mechanism is a passive response after an impact occurs, and they cannot actively avoid impacts before they occur. The suppression effect is limited by the hysteresis of the impact, and the system can still become unstable due to the impact generated during the switching process. At the same time, the optimal smoothing parameters vary greatly under different switching types and load intensities. Fixed parameters are difficult to balance the suppression effect with the dynamic response of the system. In scenarios with frequent switching or uncertain switching timing, the impact suppression effect is unstable and cannot meet the system's requirements for operational stability. Summary of the Invention

[0004] This invention provides a switching control method, system, medium, and product for a power system, which can solve the problem that it is difficult to predict and actively suppress the impact of switching when the electric-hydrogen hybrid power system switches between different behavior modes, thereby improving the stability of the power system during the behavior switching control process.

[0005] In a first aspect, an embodiment of the present invention provides a switching control method for a power system, applied to an electric-hydrogen hybrid power system, the method comprising: Collect the operating status data of the electric-hydrogen hybrid power system within a preset time period; The state change amount and state change rate are calculated using the operating state data. The operating state data, the state change amount, and the state change rate are combined to obtain a state feature sequence. The state feature sequence is input into a preset switching prediction model. The hidden layer in the switching prediction model is used to process the state feature sequence to obtain hidden state features. Each prediction branch in the switching prediction model is used to predict the hidden state features to obtain the switching probability, the switching behavior category, and the remaining switching time of the electric-hydrogen hybrid power system. If the switching probability, the switching behavior category, and the remaining switching time meet the preset switching rules, the switching behavior category is searched in a preset mode parameter mapping table to obtain the switching control parameters corresponding to the switching behavior category. Based on the remaining switching time, the control parameters to be switched, and the operating status data, a switching control strategy is generated to control the switching of the electric-hydrogen hybrid power system.

[0006] By collecting operational status data of the electric-hydrogen hybrid power system within a preset time period, a real-time and continuous data foundation can be provided for subsequent prediction of system behavior switching trends. This addresses the problem that existing technologies lack the ability to perceive changes in the system state before switching, thus failing to predict shocks in advance. This ensures the reliability of predicting shocks before switching and provides a data source guarantee for improving switching stability. By calculating the state change quantity and state change rate using operational status data and combining operational status data, state change quantity, and state change rate into a state feature sequence, the dynamic trend of system state changes can be quantified into characteristic information that can characterize the precursors of switching. This solves the problem that existing technologies only focus on the current steady-state and ignore the state changes. This paper addresses the issues of state change trends and the inability to identify handover precursors, providing precursor identification assurance to improve handover stability. By utilizing the hidden layer in the handover prediction model to process the state transition feature sequence, hidden state features are obtained. This aggregates and compresses the temporally dispersed state transition feature sequences into hidden state features containing complete temporal dependencies, ensuring that subsequent prediction branches can make accurate predictions based on sufficient temporal information, thus providing temporal feature aggregation assurance for improved handover stability. Furthermore, by using each prediction branch in the handover prediction model to predict the hidden state features, the probability of handover occurrence, the type of behavior to be handed over, and the remaining handover time can be obtained, allowing for prediction of whether a handover will occur before it actually happens. The system determines the behavior to be switched to and the remaining time before the switch, addressing the limitation of existing technologies that can only determine the likelihood of a switch but cannot simultaneously determine the switch direction and timing. This provides information integrity assurance for improved switch stability. When the probability of a switch, the type of behavior to be switched to, and the remaining time meet preset switch rules, a transition strategy is triggered, and the corresponding control parameters for the type of behavior to be switched to are retrieved from the mode parameter mapping table. This ensures that the active suppression strategy is activated only when the prediction result meets the credibility requirements, and directly maps the predicted type of behavior to be switched to an executable control objective. This addresses the problem of existing technologies lacking a triggering and determination mechanism, which leads to suppression strategies operating in uncertain situations. This application addresses the issues of blind startup in various scenarios and the difficulty in directly translating predicted results into control commands. It provides dual protection through triggering and mapping to improve switching stability. Based on the remaining switching time, control parameters to be switched, and operating status data, a switching control strategy is generated and executed. This strategy can generate a smooth transition path from the current state to the target state before the switch occurs, proactively pre-shaping energy output before the switching impact forms. It disperses the power jump and current surge at the moment of switching into the transition time window, solving the problem of existing technologies that passively respond after switching and cannot proactively avoid the impact before it forms. This suppresses the switching impact at its source, ultimately improving the stability of the power system during behavioral switching control. This application can solve the problem of difficulty in predicting and proactively suppressing the impact of switching when the electric-hydrogen hybrid power system switches between different behavioral modes, thereby improving the stability of the power system during behavioral switching control.

[0007] Furthermore, the calculation using the operational state data to obtain the state change amount and state change rate specifically includes: Several voltage changes are obtained by calculating several bus voltages of the DC bus in the operating status data. The operating status data includes several bus voltages of the DC bus, several bus currents, and several target power values ​​of the electric-hydrogen hybrid power system within the time period. By using the current of each of the busbars, several current changes of the DC busbars are obtained; The bus voltage and bus current of the DC bus at the same acquisition time are calculated to obtain several bus output powers of the DC bus within the time period, and based on the output power of each bus, several bus power changes of the DC bus are calculated. By using the target power values, several target power changes are obtained; The state change is obtained by summing up each of the voltage changes, current changes, bus power changes, and target power changes. Based on the power change of each bus and the sampling period, the state change rate is calculated, wherein the sampling period is the time interval between two adjacent acquisitions in the operating state data.

[0008] By using operational status data to calculate the state change quantity and state change rate, the dynamic trend of system state change can be quantified into characteristic information that can characterize the precursors of switching. This solves the problem that existing technologies only focus on the current steady state and ignore the state change trend, and cannot identify the precursors of switching, thus providing a precursor identification guarantee for improving switching stability.

[0009] Furthermore, the summarization of each of the voltage changes, current changes, bus power changes, and target power changes to obtain the state change specifically includes: The terminal voltage and charging / discharging current of the power battery at the same acquisition time in the operating status data are calculated to obtain several charging / discharging powers of the power battery within the time period. The electric-hydrogen hybrid power system also includes the power battery and the fuel cell. The operating status data includes several terminal voltages and several charging / discharging currents of the power battery, and several output voltages and several output currents of the fuel cell. Based on the charging and discharging power, several power power changes are calculated. The output voltage and output current of the fuel cell at the same acquisition time are calculated to obtain several fuel output powers of the fuel cell within the time period, and several fuel power changes are calculated based on each fuel output power. Using the charging and discharging power of the power battery, the fuel output power of the fuel cell, and the bus output power of the DC bus at the same acquisition time, several source-side power distribution ratios of the electric-hydrogen hybrid power system within the time period are calculated. Based on the power allocation ratios of each source side, several ratio changes are calculated. The state change is obtained by summing up each of the voltage changes, current changes, bus power changes, target power changes, power changes, fuel power changes, and proportional changes.

[0010] This further calculates the charging and discharging power of the power battery, the fuel output power of the fuel cell, and the power distribution ratio and its changes on the source side. These changes are then combined with the changes on the bus side to form a state change quantity. In addition to external electrical characteristics such as bus voltage and current, this allows for a more comprehensive characterization of the internal state evolution trend of the electric-hydrogen hybrid power system before the behavior switch, from the perspective of changes in the internal power of the energy source and the changes in the distribution relationship between the sources. This solves the problem that existing technologies, which rely solely on the electrical quantities on the bus side, cannot distinguish the contribution of the changes of different energy sources to the switching impact, resulting in insufficient characterization of state characteristics. This provides more complete information on internal state changes to improve switching stability.

[0011] Furthermore, the calculation of several source-side power allocation ratios of the electric-hydrogen hybrid power system within the time period using the charging and discharging power of the power battery, the fuel output power of the fuel cell, and the bus output power of the DC bus at the same acquisition time specifically includes: When the power battery is discharging, the charging and discharging power of the power battery and the fuel output power of the fuel cell at the same acquisition time are used to calculate several first power distribution ratios of the electric-hydrogen hybrid power system within the time period. When the power battery is charging, the fuel output power of the fuel cell and the bus output power of the DC bus at the same acquisition time are used to calculate several second power distribution ratios of the electric-hydrogen hybrid power system within the time period. By summing up the first power allocation ratios and the second power allocation ratios, a number of source-side power allocation ratios of the electric-hydrogen hybrid power system within the time period are obtained.

[0012] By employing different calculation methods for the two different operating states of the power battery—discharging and charging—the power allocation ratio can be obtained, accurately reflecting the actual power allocation relationship between the fuel cell and the power battery under different energy flow directions. This solves the problem of power allocation ratio calculation distortion caused by neglecting the differences in battery charging and discharging states in existing technologies. As a result, it provides accurate information on the allocation ratio changes for subsequent state changes, and provides more precise state change characteristics to improve switching stability.

[0013] Furthermore, if the handover probability, the type of behavior to be handed over, and the remaining handover time satisfy a preset handover rule, then the type of behavior to be handed over is retrieved from a preset mode parameter mapping table to obtain the handover control parameters corresponding to the type of behavior to be handed over, specifically including: If the switching probability, the switching behavior category, and the remaining switching time satisfy the switching rules, then the switching behavior category is used to search in the mode parameter mapping table to obtain the fuel cell switching power, power battery switching power, and power change rate switching upper limit corresponding to the switching behavior category. The switching power of the fuel cell, the switching power of the power battery, and the upper limit of the power change rate are summarized to obtain the control parameters to be switched.

[0014] In this way, when the probability of handover, the type of behavior to be handed over, and the remaining handover time meet the preset handover rules, the transition strategy is triggered and the control parameters corresponding to the type of behavior to be handed over are retrieved from the mode parameter mapping table. This ensures that the active suppression strategy is activated only when the prediction result meets the credibility requirements, and the predicted type of behavior to be handed over is directly mapped to an executable control target. This solves the problem that the suppression strategy is blindly activated in uncertain scenarios due to the lack of a triggering and judgment mechanism in the existing technology, and that the prediction result is difficult to be directly implemented into control commands. This provides a dual guarantee of triggering and mapping to improve handover stability.

[0015] Furthermore, the prediction branch includes a probability branch, a category branch, and a time branch. The step of using each prediction branch in the switching prediction model to predict the hidden state features, thereby obtaining the switching probability, the category of the behavior to be switched, and the remaining switching time of the electric-hydrogen hybrid power system, specifically includes: The hidden state features are processed using the first fully connected layer in the probability branch to obtain a first intermediate quantity, and the first intermediate quantity is normalized to obtain the switching probability. The hidden state features are processed using the second fully connected layer in the category branch to obtain the second intermediate quantity; The second intermediate quantity is normalized to obtain the behavior probabilities corresponding to several candidate behavior categories, and the behavior probabilities are filtered to determine the behavior category to be switched. The hidden state features are processed using the third fully connected layer in the time branch to obtain the third intermediate quantity; The third intermediate quantity is processed using a preset linear rectification function to obtain the remaining switching time.

[0016] By using each prediction branch in the handover prediction model to predict the hidden state features, the probability of handover, the type of behavior to be handed over, and the remaining time before handover actually occurs, it is possible to predict whether handover will occur, what behavior it will be, and how much time is left before handover actually occurs. This solves the problem that existing technologies can only determine whether handover may occur but cannot simultaneously know the direction and timing of handover, thus providing information integrity assurance for improving handover stability.

[0017] Furthermore, the step of generating a switching control strategy based on the remaining switching time, the control parameters to be switched, and the operating status data specifically includes: Based on the remaining switching time, the switching transition time for the electric-hydrogen hybrid power system to switch from the current behavior to the behavior category to be switched is calculated; Using the switching transition time as the interpolation interval, interpolation processing is performed on the control parameters to be switched and the operating status data to obtain the switching parameters corresponding to each moment within the switching transition time. The switching parameters are summarized in chronological order to obtain the switching control strategy.

[0018] This approach generates and executes a switching control strategy based on the remaining switching time, control parameters to be switched, and operating status data. It can generate a smooth transition path from the current state to the target state in advance before the switch occurs, proactively pre-shaping the energy output before the switching impact forms, and dispersing the power step and current surge at the moment of switching into the transition time window. This solves the problem that existing technologies cannot proactively avoid the impact before it forms due to passive response after the switch occurs, thereby suppressing the switching impact from the source and ultimately improving the stability of the power system in the behavioral switching control process.

[0019] Secondly, one embodiment of the present invention provides a switching control system for a power system, the switching control system being used to execute a switching control method for the power system.

[0020] Thirdly, another embodiment of the present invention provides a computer-readable storage medium, the computer-readable storage medium including a stored computer program, wherein, when the computer program is executed, it controls the device or apparatus where the computer-readable storage medium is located to perform a switching control method for a power system.

[0021] Fourthly, another embodiment of the present invention provides a computer program product, including a computer program or instructions, which, when executed by a communication device, implements a switching control method for a power system. Attached Figure Description

[0022] To more clearly illustrate the technical solution of this application, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0023] Figure 1 This is a flowchart illustrating one embodiment of the switching control method for a power system provided in this application; Figure 2 This is a flowchart illustrating steps S201 to S206 provided in this application; Figure 3 This is a flowchart illustrating steps S301 to S306 provided in this application. Detailed Implementation

[0024] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below with reference to the accompanying drawings of the embodiments. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0025] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains; the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the application; the terms “comprising” and “having”, and any variations thereof, in the specification, claims, and foregoing description of the drawings are intended to cover non-exclusive inclusion.

[0026] In the description of the embodiments of this application, technical terms such as "first" and "second" are used only to distinguish different objects and should not be construed as indicating or implying relative importance or implicitly specifying the number, specific order, or primary and secondary relationship of the indicated technical features. In the description of the embodiments of this application, "multiple" means two or more, unless otherwise explicitly defined.

[0027] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a mutually exclusive, independent, or alternative embodiment. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0028] In the description of the embodiments in this application, the term "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Additionally, the character " / " in this document generally indicates that the preceding and following related objects have an "or" relationship.

[0029] In the description of the embodiments of this application, the term "multiple" refers to two or more (including two), similarly, "multiple sets" refers to two or more (including two sets), and "multiple pieces" refers to two or more (including two pieces).

[0030] In the description of the embodiments of this application, unless otherwise expressly specified and limited, technical terms such as "installation," "connection," "joining," and "fixing" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components. Those skilled in the art can understand the specific meaning of the above terms in the embodiments of this application according to the specific circumstances.

[0031] In the field of power system energy management technology, when switching between different behavior modes, the power reference value and energy distribution ratio of an electric-hydrogen hybrid power system are prone to sudden changes, which can cause energy surges, trigger protection actions, and reduce system lifespan. Existing technologies use filtering and slope limiting for smooth control, but these are passive responses and cannot actively avoid surges. Furthermore, fixed parameters are difficult to adapt to different switching scenarios, resulting in unstable surge suppression effects and failing to meet operational stability requirements.

[0032] See Figure 1In order to solve the problem that it is difficult to predict and actively suppress the impact of switching when the electric hydrogen hybrid power system switches between different behavior modes, so as to improve the stability of the power system in the behavior switching control process, an embodiment of the present invention provides a switching control method for a power system, including steps S101 to S103. Step S101: Collect the operating status data of the electric-hydrogen hybrid power system within a preset time period; In some embodiments, the operation status data of the electric-hydrogen hybrid power system is collected within a preset time period, specifically including: collecting system operation status data in the controller at a fixed sampling period (e.g., 10ms), and establishing a sliding window (e.g., with a length of 50) to store the most recent (e.g., the most recent 0.5s) operation status sequence; at the arrival of each sampling period, the collected raw data includes: DC bus voltage sensor value, DC bus current sensor value, fuel cell output voltage and output current, power battery terminal voltage and charging / discharging current (positive value for discharging, negative value for charging), and external command signals issued by the upper-level controller (including target power value and task mode command); storing these raw data into the sliding window, while deleting the earliest data in the window to keep the window length unchanged; organizing all state quantities into a state vector in a fixed order, with each sampling point containing a state vector, and obtaining a continuous state sequence after collection as the system's operation status data.

[0033] For example, after collecting the above data, other data can also be collected, including: key actuator state variables (such as the duty cycle of the DC / DC converter and the speed of the fuel cell air compressor) are read from the internal register of the controller; the state of charge (SOC) of the power battery is estimated by the ampere-hour integration method or Kalman filtering.

[0034] Step S102: Calculate the state change amount and state change rate using the operating state data. Combine the operating state data, the state change amount, and the state change rate to obtain a state feature sequence. Input the state feature sequence into a preset switching prediction model. Process the state feature sequence using the hidden layer in the switching prediction model to obtain hidden state features. Predict the hidden state features using each prediction branch in the switching prediction model to obtain the switching probability, the switching behavior category, and the remaining switching time of the electric-hydrogen hybrid power system. If the switching probability, the switching behavior category, and the remaining switching time satisfy a preset switching rule, retrieve the switching control parameters corresponding to the switching behavior category from a preset mode parameter mapping table using the switching behavior category. See Figure 2In some embodiments, the calculation of the state change amount and state change rate using the running state data specifically includes steps S201 to S206. Step S201: Calculate several voltage changes using several bus voltages of the DC bus in the operating status data. The operating status data includes several bus voltages, several bus currents, and several target power values ​​of the electric-hydrogen hybrid power system within the time period. In some embodiments, several voltage changes are calculated using several bus voltages of the DC bus in the operating status data. The operating status data includes several bus voltages, several bus currents, and several target power values ​​of the electric-hydrogen hybrid power system within the time period. Specifically, this includes: performing a difference operation on the bus voltage values ​​of each pair of adjacent sampling points within the sliding window; subtracting the bus voltage value of the previous sampling time from the bus voltage value of the later sampling time to obtain the voltage change corresponding to that pair of adjacent sampling points; and iterating through all pairs of adjacent sampling points within the sliding window to obtain several voltage changes. The operating status data includes the bus voltage, bus current, and target power value of the electric-hydrogen hybrid power system corresponding to each moment within the sampling period.

[0035] In some embodiments, calculations are performed using several bus voltages of the DC bus in the operating status data to obtain relevant formulas for several voltage changes, specifically including: Formula for calculating voltage change: ; In the formula, This represents the voltage change at the DC bus. Let be the bus voltage of the DC bus at time t.

[0036] Step S202: Calculate using the current of each bus to obtain several current changes of the DC bus. In some embodiments, the calculation is performed using the bus currents to obtain several current changes of the DC bus, specifically including: performing a difference operation on the bus current values ​​of each pair of adjacent sampling points within the sliding window, subtracting the bus current value of the previous sampling time from the bus current value of the later sampling time to obtain the current change corresponding to the pair of adjacent sampling points, and traversing all pairs of adjacent sampling points within the sliding window to obtain several current changes.

[0037] In some embodiments, calculations are performed using the currents of each bus to obtain relevant formulas for several current changes of the DC bus, specifically including: Formula for calculating the change in current: ; In the formula, This represents the change in current at the DC bus. Let be the bus current of the DC bus at time t.

[0038] Step S203: Calculate the bus voltage and bus current of the DC bus at the same acquisition time to obtain several bus output powers of the DC bus within the time period, and calculate several bus power changes of the DC bus based on each bus output power. In some embodiments, the bus voltage and bus current of the DC bus at the same acquisition time are calculated to obtain several bus output powers of the DC bus within the time period, and several bus power changes of the DC bus are calculated based on each bus output power. Specifically, this includes: multiplying the bus voltage and bus current at the same acquisition time to obtain the bus output power at that time; traversing each acquisition time within the sliding window to obtain several bus output powers; and performing a difference operation on the bus output power values ​​of each pair of adjacent sampling points in the bus output power sequence to obtain the bus power change corresponding to each pair of adjacent sampling points.

[0039] In some embodiments, the bus voltage and bus current of the DC bus at the same acquisition time are calculated to obtain several bus output powers of the DC bus within the time period, and based on each bus output power, relevant formulas for the changes in several bus power of the DC bus are calculated, specifically including: Formula for calculating bus output power: ; Formula for calculating bus power change: ; In the formula, This refers to the bus voltage. Bus current; The DC bus output power at time t; This represents the change in bus power.

[0040] Step S204: Calculate using each of the target power values ​​to obtain several target power changes; In some embodiments, several target power changes are calculated using the target power values, specifically including: performing a difference operation on the target power values ​​of each pair of adjacent sampling points within the sliding window, subtracting the target power value of the previous sampling time from the target power value of the later sampling time to obtain the target power change corresponding to the pair of adjacent sampling points, and traversing all pairs of adjacent sampling points within the sliding window to obtain several target power changes.

[0041] In some embodiments, calculations are performed using the target power values ​​to obtain relevant formulas for several target power changes, specifically including: Formula for calculating the change in target power: ; In the formula, The target power value of the electric-hydrogen hybrid power system at time t; This represents the change in target power.

[0042] Step S205: Summarize the voltage changes, current changes, bus power changes, and target power changes to obtain the state change amount; See Figure 3 In some embodiments, the step of summing up each of the voltage changes, each of the current changes, each of the bus power changes and each of the target power changes to obtain the state change amount specifically includes steps S301 to S306. Step S301: Calculate the terminal voltage and charging / discharging current of the power battery at the same acquisition time in the operating status data to obtain several charging / discharging powers of the power battery within the time period. The electric-hydrogen hybrid power system also includes the power battery and the fuel cell. The operating status data includes several terminal voltages and several charging / discharging currents of the power battery, and several output voltages and several output currents of the fuel cell. In some embodiments, the terminal voltage and charging / discharging current of the power battery at the same acquisition time in the operating status data are calculated to obtain several charging / discharging powers of the power battery within the time period. The electric-hydrogen hybrid power system further includes the power battery and the fuel cell. The operating status data includes several terminal voltages and several charging / discharging currents of the power battery, and several output voltages and several output currents of the fuel cell. Specifically, this includes multiplying the terminal voltage of the power battery at the same acquisition time by the charging / discharging current to obtain the charging / discharging power of the power battery at that time, wherein the sign of the charging / discharging current indicates the direction (positive value for discharging, negative value for charging). The power battery charging / discharging power is obtained by traversing each acquisition time within the sliding window.

[0043] In some embodiments, the terminal voltage and charging / discharging current of the power battery at the same acquisition time in the operating status data are calculated to obtain relevant formulas for several charging / discharging powers of the power battery within the time period, specifically including: Formula for calculating charging and discharging power: = × ; In the formula, Terminal voltage; This refers to the charging and discharging current. This refers to the charging and discharging power.

[0044] Step S302: Based on the charging and discharging power, calculate several power power changes; In some embodiments, based on each of the charging and discharging powers, several power power changes are calculated, specifically including: performing a difference operation on the charging and discharging power values ​​of each pair of adjacent sampling points in the power battery charging and discharging power sequence, subtracting the charging and discharging power of the previous sampling point from the charging and discharging power of the later sampling time to obtain the power power change, and traversing all pairs of adjacent sampling points to obtain several power power changes.

[0045] In some embodiments, based on the charging and discharging power, several formulas related to the changes in power are calculated, specifically including: Formula for calculating the change in power output: ; In the formula, This refers to the change in power output; This represents the charging and discharging power of the power battery at time t.

[0046] It should be noted that a battery charging / discharging direction change indicator can be set, specifically: if ≥0 and <0, then the direction change flag =1, otherwise 0.

[0047] Step S303: Calculate the output voltage and output current of the fuel cell at the same acquisition time to obtain several fuel output powers of the fuel cell within the time period, and calculate several fuel power changes based on each fuel output power. In some embodiments, the output voltage and output current of the fuel cell at the same sampling time are calculated to obtain several fuel output powers of the fuel cell within the time period, and several fuel power changes are calculated based on each fuel output power. Specifically, this includes: multiplying the fuel cell output voltage and output current at the same sampling time to obtain the fuel cell output power at that time; traversing each sampling time within the sliding window to obtain several fuel cell output powers; and performing a difference operation on the fuel cell output power values ​​of each pair of adjacent sampling points in the fuel cell output power sequence to obtain the fuel power change.

[0048] In some embodiments, the output voltage and output current of the fuel cell at the same acquisition time are calculated to obtain several fuel output powers of the fuel cell within the time period, and based on each fuel output power, relevant formulas for several fuel power changes are calculated, specifically including: Formula for calculating fuel power variation: ; In the formula, This represents the change in fuel power. Let be the fuel output power of the fuel cell at time t.

[0049] Step S304: Using the charging and discharging power of the power battery, the fuel output power of the fuel cell, and the bus output power of the DC bus at the same acquisition time, calculate several source-side power distribution ratios of the electric-hydrogen hybrid power system within the time period. In some embodiments, calculating several source-side power allocation ratios of the electric-hydrogen hybrid power system within the time period using the charging / discharging power of the power battery, the fuel output power of the fuel cell, and the bus output power of the DC bus at the same acquisition time specifically includes: when the power battery is discharging, calculating several first power allocation ratios of the electric-hydrogen hybrid power system within the time period using the charging / discharging power of the power battery and the fuel output power of the fuel cell at the same acquisition time; when the power battery is charging, calculating several second power allocation ratios of the electric-hydrogen hybrid power system within the time period using the fuel output power of the fuel cell and the bus output power of the DC bus at the same acquisition time; and summing up the first power allocation ratios and the second power allocation ratios to obtain several source-side power allocation ratios of the electric-hydrogen hybrid power system within the time period. Specifically, the charging and discharging state of the power battery is determined. When the charging and discharging current is greater than zero, it is determined to be in a discharging state; when the charging and discharging current is less than zero, it is determined to be in a charging state. In the discharging state, the fuel cell output power is divided by the sum of the absolute values ​​of the fuel cell output power and the charging and discharging power of the power battery to obtain the first power allocation ratio. In the charging state, the fuel cell output power is divided by the DC bus output power to obtain the second power allocation ratio. The above method is used to traverse each acquisition time to obtain the power allocation ratio corresponding to each time, forming a source-side power allocation ratio sequence.

[0050] In some embodiments, the formula for calculating the power distribution ratio of the electric-hydrogen hybrid power system within the time period using the charging and discharging power of the power battery, the fuel output power of the fuel cell, and the bus output power of the DC bus at the same acquisition time specifically includes: The formula for calculating the first power allocation ratio under discharge conditions is as follows: ; The formula for calculating the second power allocation ratio during charging is as follows: ; In the formula, This is the first power allocation ratio; This is the second power allocation ratio; This refers to the fuel output power of the fuel cell; The charging and discharging power of the power battery; This refers to the charging and discharging power of the DC bus.

[0051] By employing different calculation methods for the two different operating states of the power battery—discharging and charging—the power allocation ratio can be obtained, accurately reflecting the actual power allocation relationship between the fuel cell and the power battery under different energy flow directions. This solves the problem of power allocation ratio calculation distortion caused by neglecting the differences in battery charging and discharging states in existing technologies. As a result, it provides accurate information on the allocation ratio changes for subsequent state changes, and provides more precise state change characteristics to improve switching stability.

[0052] Step S305: Based on the power allocation ratio of each source side, calculate several ratio changes; In some embodiments, based on each of the source-side power allocation ratios, several ratio changes are calculated, specifically including: performing a difference operation on the power allocation ratio values ​​of each pair of adjacent sampling points in the source-side power allocation ratio sequence, subtracting the power allocation ratio of the previous sampling time from the power allocation ratio of the later sampling time to obtain the ratio change, and traversing all adjacent sampling point pairs to obtain several ratio changes.

[0053] In some embodiments, based on the respective source-side power allocation ratios, formulas are used to calculate several proportional changes, specifically including: Formula for calculating proportional change: ; In the formula, It represents the proportional change; Let be the source-side power allocation ratio at time t.

[0054] Step S306: Summarize the voltage changes, current changes, bus power changes, target power changes, power changes, fuel power changes, and proportional changes to obtain the state change.

[0055] In some embodiments, the voltage changes, current changes, bus power changes, target power changes, power changes, fuel power changes, and proportional changes are summarized to obtain the state change amount. Specifically, this includes aligning the above-mentioned changes in time order, arranging them in a preset fixed order, and merging them into a sequence of multidimensional state change amounts.

[0056] This further calculates the charging and discharging power of the power battery, the fuel output power of the fuel cell, and the power distribution ratio and its changes on the source side. These changes are then combined with the changes on the bus side to form a state change quantity. In addition to external electrical characteristics such as bus voltage and current, this allows for a more comprehensive characterization of the internal state evolution trend of the electric-hydrogen hybrid power system before the behavior switch, from the perspective of changes in the internal power of the energy source and the changes in the distribution relationship between the sources. This solves the problem that existing technologies, which rely solely on the electrical quantities on the bus side, cannot distinguish the contribution of the changes of different energy sources to the switching impact, resulting in insufficient characterization of state characteristics. This provides more complete information on internal state changes to improve switching stability.

[0057] Step S206: Based on the power change of each bus and the sampling period, calculate the state change rate, wherein the sampling period is the time interval between two adjacent acquisitions in the operating state data.

[0058] In some embodiments, the state change rate is calculated based on the power change of each bus and the sampling period, wherein the sampling period is the time interval between two adjacent acquisitions in the operating state data, specifically including: dividing each power change of the bus by the sampling period to obtain the power change rate of the bus corresponding to each pair of adjacent sampling points.

[0059] In some embodiments, the state change rate is calculated based on the power change of each bus and the sampling period, wherein the sampling period is the time interval between two adjacent acquisitions of the operating state data, specifically including: Formula for calculating the rate of change of state: ; In the formula, The rate of change of state; This represents the change in bus power. The sampling period.

[0060] By using operational status data to calculate the state change quantity and state change rate, the dynamic trend of system state change can be quantified into characteristic information that can characterize the precursors of switching. This solves the problem that existing technologies only focus on the current steady state and ignore the state change trend, and cannot identify the precursors of switching, thus providing a precursor identification guarantee for improving switching stability.

[0061] For example, this application will also collect the current behavior label of the system, specifically including: the current behavior label is obtained by judging the rules pre-stored in the controller. If |target power value - DC bus power value| is less than a preset threshold and SOC is in the target range and fuel cell power change rate is less than the threshold, it is judged as steady-state power supply behavior; if the target power value increases by a step and SOC is greater than the lower limit, it is judged as peak power compensation behavior; if the target power value decreases by a step or the load braking feedback signal is valid, it is judged as regenerative braking behavior; external context signals directly read changes in upper-level task instructions (such as switching from cruise to rapid acceleration), operator throttle / brake opening change rate, load level changes (such as hill climbing gradient sensor signal), etc.; this label can be used to determine the behavior category currently executed by the system.

[0062] In some embodiments, the running state data, the state change amount, and the state change rate are combined to obtain a state feature sequence. Specifically, this includes: arranging the running state data, state change amount, and state change rate at each sampling time within the sliding window in a preset fixed order to form a state feature vector corresponding to each sampling time; and arranging the state feature vectors at all sampling times in chronological order to obtain a state feature sequence.

[0063] In some embodiments, the training process of the switching prediction model specifically includes: the switching prediction model adopts a gated recurrent unit network (GRU), the input dimension is equal to the dimension of the state feature sequence, the number of hidden layer neurons is set to 64, and the output layer contains three branches; the training of the model is divided into two parts: offline training and online incremental training; during offline training, data on various switching events that occur in the actual operation of the electric-hydrogen hybrid power system are collected, and the switching probability, remaining switching time, and the category of the behavior to be switched (one-hot encoding) at each moment are manually labeled. The state feature sequence is used as input, and the above three labels are used as output. The remaining switching time branch is trained using the mean squared error loss function, and the switching probability branch and the category of the behavior to be switched are trained using the cross-entropy loss function. The Adam optimizer is used for training, and the basic weights are fixed after training; during online incremental training, after each switching in the actual operation of the system, the feature sequence before the current switching and the real switching label are used as a new sample, and the model is updated with a small batch gradient descent (the learning rate is set to 0.001) so that the model gradually adapts to system aging or operating condition drift.

[0064] It should be noted that the Gated Recurrent Unit (GRU) is a variant of the recurrent neural network that controls the updating and forgetting of information through a gating mechanism, and can effectively capture long-term dependencies in time series; the Adam optimizer is a stochastic gradient descent optimization algorithm with an adaptive learning rate.

[0065] In some embodiments, the state feature sequence is input into a preset switching prediction model to process the state feature sequence using the hidden layer in the switching prediction model to obtain hidden state features. Specifically, this includes: inputting each feature vector in the state feature sequence into the GRU hidden layer of the switching prediction model step by step; updating the current hidden state of the GRU hidden layer at each time step based on the currently input feature vector and the hidden state of the previous time step; and accumulating all temporal information from the beginning of the sequence to the current time step through iteration step by step. After all time steps are processed, the hidden state of the last time step is taken as the hidden state feature of the entire state feature sequence.

[0066] In some embodiments, the state feature sequence is input into a preset switching prediction model to process the state feature sequence using a hidden layer in the switching prediction model, thereby obtaining relevant formulas for the hidden state features, specifically including: The mathematical expression for the hidden layer: ; In the formula, The state feature vector input at the current time step; This is the hidden state from the previous time step; This is the hidden state after the current time step update.

[0067] In some embodiments, the prediction branch includes a probability branch, a category branch, and a time branch. The step of using each prediction branch in the switching prediction model to predict the hidden state features and obtain the switching probability, the type of behavior to be switched, and the remaining switching time of the electric-hydrogen hybrid power system specifically includes: processing the hidden state features using a first fully connected layer in the probability branch to obtain a first intermediate quantity, and normalizing the first intermediate quantity to obtain the switching probability; processing the hidden state features using a second fully connected layer in the category branch to obtain a second intermediate quantity; normalizing the second intermediate quantity to obtain behavior probabilities corresponding to several candidate behavior types, and filtering each behavior probability to determine the type of behavior to be switched; processing the hidden state features using a third fully connected layer in the time branch to obtain a third intermediate quantity; and processing the third intermediate quantity using a preset linear rectification function to obtain the remaining switching time. Specifically, the hidden state features are input into the first fully connected layer of the probability branch. The first fully connected layer (input dimension 64, output dimension 2) performs a linear transformation on the hidden state features, mapping the high-dimensional hidden state features into a two-dimensional vector. The two elements of this two-dimensional vector correspond to the unnormalized scores for switching and not switching, respectively, obtaining the first intermediate value. The first intermediate value is then subjected to Softmax normalization, converting the two unnormalized scores into two probability values. The sum of the two probability values ​​is 1, and the probability value corresponding to switching is taken as the probability of switching occurring. The hidden state features are then input into the second fully connected layer of the category branch. The second fully connected layer (input dimension 6, output dimension K, where K is the total number of behavior categories) performs a linear transformation on the hidden state features, mapping the high-dimensional hidden state features into a two-dimensional vector. The first intermediate value is a K-dimensional vector, where each element corresponds to the unnormalized score of each candidate behavior category. Softmax normalization is applied to this second intermediate value, converting each unnormalized score into a behavior probability corresponding to each candidate behavior category. The sum of all behavior probabilities is 1. The behavior category corresponding to the maximum value among these probabilities is selected as the behavior category to be switched. The hidden state features are input into the third fully connected layer of the time branch. This third fully connected layer (64 input dimensions, 1 output dimension) performs a linear transformation on the hidden state features, mapping the high-dimensional hidden state features to a one-dimensional scalar. This scalar represents the unconstrained predicted time value, resulting in the third intermediate value. A Rectified Linear Unit (ReLU) function is applied to the third intermediate value, setting negative values ​​to zero and keeping positive values ​​unchanged to ensure that the output remaining switching time is non-negative.

[0068] By using each prediction branch in the handover prediction model to predict the hidden state features, the probability of handover, the type of behavior to be handed over, and the remaining time before handover actually occurs, it is possible to predict whether handover will occur, what behavior it will be, and how much time is left before handover actually occurs. This solves the problem that existing technologies can only determine whether handover may occur but cannot simultaneously know the direction and timing of handover, thus providing information integrity assurance for improving handover stability.

[0069] In some embodiments, if the switching probability, the type of behavior to be switched, and the remaining switching time satisfy a preset switching rule, then the type of behavior to be switched is used to search a preset mode parameter mapping table to obtain the switching control parameters corresponding to the type of behavior to be switched. Specifically, this includes: if the switching probability, the type of behavior to be switched, and the remaining switching time satisfy the switching rule, then the type of behavior to be switched is used to search the mode parameter mapping table to obtain the fuel cell switching power, power battery switching power, and power change rate switching upper limit corresponding to the type of behavior to be switched; and then the fuel cell switching power, power battery switching power, and power change rate switching upper limit are summarized to obtain the switching control parameters. Specifically, it determines whether the probability of switching occurs reaches a preset probability threshold (e.g., 0.75). If so, it further determines whether the remaining switching time falls within a preset lead time threshold range (e.g., 0.05 seconds to 0.5 seconds). If so, it further determines whether the type of behavior to be switched remains consistent within several consecutive sampling periods (e.g., 3). If so, it determines that switching is about to occur and triggers the transition strategy generation process. After triggering, it enters the transition execution state, in which new transition strategies are not repeatedly triggered. After triggering, using the type of behavior to be switched as an index, it looks up the corresponding fuel cell switching power, power battery switching power, and power change rate switching upper limit in the pre-stored mode-parameter mapping table and summarizes them into the control parameters to be switched.

[0070] For example, when the behavior to be switched is a steady-state high-efficiency energy supply behavior, the switching power of the fuel cell is 0.8 times the target total output power, the switching power of the power battery is 0.2 times the target total output power, and the upper limit of the power change rate is 2kW / s; when the behavior to be switched is a peak power compensation behavior, the switching power of the fuel cell is 0.3 times the target total output power, the switching power of the power battery is 0.7 times the target total output power, and the upper limit of the power change rate is 5kW / s; when the behavior to be switched is a regenerative braking behavior, the switching power of the fuel cell is 0.5 times the target total output power, the switching power of the power battery is negative (i.e., charging power), and the upper limit of the power change rate is 1kW / s.

[0071] It should be noted that the above-mentioned fuel cell switching power refers to the target power value that the fuel cell should switch to under the target behavior, the power battery switching power refers to the target power value that the power battery should switch to under the target behavior, and the upper limit of power change rate refers to the highest allowable value of the change rate of each power parameter during the transition process.

[0072] In this way, when the probability of handover, the type of behavior to be handed over, and the remaining handover time meet the preset handover rules, the transition strategy is triggered and the control parameters corresponding to the type of behavior to be handed over are retrieved from the mode parameter mapping table. This ensures that the active suppression strategy is activated only when the prediction result meets the credibility requirements, and the predicted type of behavior to be handed over is directly mapped to an executable control target. This solves the problem that the suppression strategy is blindly activated in uncertain scenarios due to the lack of a triggering and judgment mechanism in the existing technology, and that the prediction result is difficult to be directly implemented into control commands. This provides a dual guarantee of triggering and mapping to improve handover stability.

[0073] Step S103: Based on the remaining switching time, the control parameters to be switched, and the operating status data, a switching control strategy is generated to control the switching of the electric-hydrogen hybrid power system. In some embodiments, generating a switching control strategy based on the remaining switching time, the control parameters to be switched, and the operating status data specifically includes: calculating the switching transition time for the electric-hydrogen hybrid power system to switch from its current behavior to the behavior category to be switched based on the remaining switching time; interpolating the control parameters to be switched and the operating status data using the switching transition time as an interpolation interval to obtain the switching parameters corresponding to each moment within the switching transition time; and summarizing the switching parameters in chronological order to obtain the switching control strategy. Specifically, the initial value of the switching transition time is determined based on the remaining switching time. The remaining switching time is multiplied by the transition weight (e.g., 0.8, to reserve a 20% margin to ensure the transition is completed before the switch arrives), and then limited to a preset time interval (e.g., between 0.1 seconds and 0.5 seconds). The control reference value at the current moment is read as the starting point (including the current fuel cell power, current power battery power, and current power change rate limit), and the switching power of the fuel cell, power battery, and power change rate upper limit in the control parameters to be switched are used as the ending point. During the switching transition time, for the fuel cell power and power battery power, a linear ramp interpolation method is used to uniformly change from the current value to the target switching value. That is, in each sampling period, the current value is increased or decreased by the ratio of the total difference to the total number of steps until the target value is reached, forming the switching parameters corresponding to each sampling moment within the transition duration. The switching parameters at each moment are arranged in chronological order to obtain the switching control strategy.

[0074] In some embodiments, generating relevant formulas for the handover control strategy based on the remaining handover time, the control parameters to be handed over, and the operating status data specifically includes: The initial calculation formula for the transition duration is as follows: ; The calculation formula for linear ramp interpolation (taking power as an example): ; ; In the formula, To switch to the remaining time; For transition weights; and These are the values ​​of the left and right endpoints of the preset time interval, respectively. The current value; Switch the target value; This is the current step number; This represents the total number of steps. The value to be switched for the current step number; The duration of the transition; The sampling period.

[0075] For example, for an electric-hydrogen hybrid power system, during the transition strategy execution, the power of the fuel cell and the power battery can be adjusted differentially according to different target behavior categories. Specifically, when the behavior category to be switched is steady-state high-efficiency energy supply behavior, the target power ratio of the fuel cell can be determined according to the optimal efficiency range of the fuel cell (e.g., the fuel cell is most efficient at 40kW). If the current fuel cell power ratio is 0.5 and the target fuel cell power ratio is 0.8, then during the transition period, the fuel cell power will be linearly increased from the current value to 0.8 times the target total output power, while the power battery power will be linearly decreased from the current value to 0.2 times the target total output power. The step size is determined by the ratio of the difference to the total number of steps. When the behavior to be switched is peak power compensation, the target power of the power battery is the smaller of the difference between the load demand power and the maximum output power of the fuel cell and the maximum discharge power of the power battery. If the current power battery power is zero and the target value is 30kW, it increases linearly. At the same time, the target power of the fuel cell adopts a slower rate of change, and the target value is 80% of the maximum output power of the fuel cell. When the behavior to be switched is regenerative braking, the target total injected power is the load demand power minus the feedback power. The fuel cell power is reduced first, and the remaining power is absorbed by the power battery (a negative value indicates charging). The adjustment step size also adopts a linear ramp method.

[0076] For example, for an electric-hydrogen hybrid power system, during the execution of the transition strategy, relevant formulas can be used to differentiate the power of the fuel cell and the power battery according to different target behavior categories. These formulas specifically include: Formula for calculating the target power of the power battery during peak power compensation behavior: ; Formula for calculating the total injected power of the target in regenerative braking behavior: ; In the formula, Power required by the load; This represents the maximum output power of the fuel cell; This refers to the maximum discharge power of the power battery. For feedback power; Total power injected to the target.

[0077] For example, before generating the transition sequence, the transition time can be corrected based on the current bus voltage deviation, bus current margin, load strength level, and energy storage unit status. Specifically, if the current bus voltage is lower than the rated value (e.g., 450V for a 500V system) and the voltage deviation exceeds a preset threshold (e.g., 20V), the transition time will be extended. If the bus current margin (i.e., the difference between the maximum allowable current and the current bus current divided by the maximum allowable current) is large and the load strength level is low, the transition time can be appropriately shortened. If the power battery SOC is lower than 20%, the transition time for transitions requiring deep battery discharge will be forcibly extended and the target discharge power limit will be reduced. If the energy storage unit direction change indicator indicates a change from discharging to charging, the transition time will be extended to prevent current reverse impact. The corrected transition time will then recheck the power change rate constraint to ensure that the change rate does not exceed the power change rate switching limit in the control parameters to be switched.

[0078] For example, before generating the transition sequence, relevant formulas can be used to correct the transition time based on the current bus voltage deviation, bus current margin, load strength level, and energy storage unit status. These formulas specifically include: The expression for the power change rate constraint: ; In the formula, The current value; Switch the target value; The duration of the transition; Set the upper limit for the power change rate.

[0079] It should be noted that if the power change rate constraint does not hold, the transition time needs to be extended further until the constraint condition is met.

[0080] This approach generates and executes a switching control strategy based on the remaining switching time, control parameters to be switched, and operating status data. It can generate a smooth transition path from the current state to the target state in advance before the switch occurs, proactively pre-shaping the energy output before the switching impact forms, and dispersing the power step and current surge at the moment of switching into the transition time window. This solves the problem that existing technologies cannot proactively avoid the impact before it forms due to passive response after the switch occurs, thereby suppressing the switching impact from the source and ultimately improving the stability of the power system in the behavioral switching control process.

[0081] In some embodiments, the switching control strategy is used to switch the electric-hydrogen hybrid power system, specifically including: reading the switching parameters corresponding to the current moment from the switching control strategy, sending the fuel cell power value in the switching parameters to the fuel cell DC / DC controller, sending the power battery power value to the power battery bidirectional DC / DC controller, sending the power change rate limit value to the corresponding power management module, and having each controller perform the corresponding power adjustment action according to the received target value.

[0082] In some embodiments, after switching the electric-hydrogen hybrid power system using the switching control strategy, the method further includes: after the switching is completed, quantitatively evaluating the impact suppression effect and updating the threshold and prediction model parameters accordingly; the switching completion determination condition is that within 5 consecutive sampling periods, the category of the behavior to be switched remains unchanged, and the deviation between the current control reference and the control reference to be switched is less than a preset error, thus determining that the switching is completed; after determining that the switching is completed, calculating the impact indicators within a preset evaluation time window (e.g., 0.2s before and after the switching occurs, for a total of 0.4s), including: power step amplitude, peak power change rate, peak bus voltage offset, and peak bus current, and calculating the comprehensive impact score using a normalized weighted method; when the comprehensive impact score exceeds 0.6, it is determined that the suppression is insufficient, and an update is performed, reducing the switching probability threshold from the current value by 0.05 (but not lower than 0.6), or expanding the allowable range of the remaining switching data (lower limit reduced by 0.02s, upper limit increased by 0.05s), using the data before this switching. The feature sequence is used to perform incremental training on the prediction model. If the impact is mainly manifested as a bus voltage drop, the transition time correction coefficient for subsequent similar switching scenarios is increased (by 0.1). If the impact is mainly manifested as a current surge, the upper limit of the power change rate for the corresponding target behavior category is reduced by 10%. When the transition strategy is triggered but no actual switching occurs within the preset observation window (maximum transition duration after triggering + 0.2s), it is judged as a false alarm. At this time, the trigger threshold is increased (the probability threshold is increased by 0.05, not exceeding 0.95, or the target behavior stability judgment period is extended from 3 sampling periods to 5) to reduce unnecessary transition shaping. When the transition strategy is not triggered but the system actually undergoes a significant behavior switch within the subsequent short time window and the impact score exceeds 0.6, it is judged as a missed alarm. At this time, the prediction probability threshold is reduced (by 0.05) or the allowed advance trigger range is expanded, and this sample is written into the incremental training set as a key sample. After completing the threshold and model update, the system exits the transition execution state and returns to the waiting-to-trigger state.

[0083] In some embodiments, the relevant formulas after switching control of the electric-hydrogen hybrid power system using the switching control strategy further include: Power step amplitude: ; Peak power change rate: ; Peak bus voltage offset: ; Peak bus current: ; Overall Impact Rating: ; In the formula, This refers to the output power after switching. Output power before switching; To switch the actual time of occurrence (determined by the behavior label jump point); This represents the change in bus power. The sampling period; This refers to the bus voltage. This is the rated value of the bus voltage; Bus current; The power change rate safety threshold; Voltage safety threshold; The current safety threshold; 、 、 and These are the weighting coefficients; This is the power safety threshold.

[0084] By collecting operational status data of the electric-hydrogen hybrid power system within a preset time period, a real-time and continuous data foundation can be provided for subsequent prediction of system behavior switching trends. This addresses the problem that existing technologies lack the ability to perceive changes in the system state before switching, thus failing to predict shocks in advance. This ensures the reliability of predicting shocks before switching and provides a data source guarantee for improving switching stability. By calculating the state change quantity and state change rate using operational status data and combining operational status data, state change quantity, and state change rate into a state feature sequence, the dynamic trend of system state changes can be quantified into characteristic information that can characterize the precursors of switching. This solves the problem that existing technologies only focus on the current steady-state and ignore the state changes. This paper addresses the issues of state change trends and the inability to identify handover precursors, providing precursor identification assurance to improve handover stability. By utilizing the hidden layer in the handover prediction model to process the state transition feature sequence, hidden state features are obtained. This aggregates and compresses the temporally dispersed state transition feature sequences into hidden state features containing complete temporal dependencies, ensuring that subsequent prediction branches can make accurate predictions based on sufficient temporal information, thus providing temporal feature aggregation assurance for improved handover stability. Furthermore, by using each prediction branch in the handover prediction model to predict the hidden state features, the probability of handover occurrence, the type of behavior to be handed over, and the remaining handover time can be obtained, allowing for prediction of whether a handover will occur before it actually happens. The system determines the behavior to be switched to and the remaining time before the switch, addressing the limitation of existing technologies that can only determine the likelihood of a switch but cannot simultaneously determine the switch direction and timing. This provides information integrity assurance for improved switch stability. When the probability of a switch, the type of behavior to be switched to, and the remaining time meet preset switch rules, a transition strategy is triggered, and the corresponding control parameters for the type of behavior to be switched to are retrieved from the mode parameter mapping table. This ensures that the active suppression strategy is activated only when the prediction result meets the credibility requirements, and directly maps the predicted type of behavior to be switched to an executable control objective. This addresses the problem of existing technologies lacking a triggering and determination mechanism, which leads to suppression strategies operating in uncertain situations. This application addresses the issues of blind startup in various scenarios and the difficulty in directly translating predicted results into control commands. It provides dual protection through triggering and mapping to improve switching stability. Based on the remaining switching time, control parameters to be switched, and operating status data, a switching control strategy is generated and executed. This strategy can generate a smooth transition path from the current state to the target state before the switch occurs, proactively pre-shaping energy output before the switching impact forms. It disperses the power jump and current surge at the moment of switching into the transition time window, solving the problem of existing technologies that passively respond after switching and cannot proactively avoid the impact before it forms. This suppresses the switching impact at its source, ultimately improving the stability of the power system during behavioral switching control. This application can solve the problem of difficulty in predicting and proactively suppressing the impact of switching when the electric-hydrogen hybrid power system switches between different behavioral modes, thereby improving the stability of the power system during behavioral switching control.

[0085] Based on the above-described embodiment of the switching control method for a power system, an embodiment of the present invention provides a switching control system for a power system, wherein the switching control system is used to execute the switching control method for the power system.

[0086] Based on the above-described embodiment of the power system switching control method, another embodiment of the present invention provides a terminal device, which includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements a power system switching control method according to any embodiment of the present invention.

[0087] For example, in this embodiment, the computer program can be divided into one or more modules, which are stored in the memory and executed by the processor to complete the present invention. The one or more modules may be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of the computer program in the terminal device.

[0088] The terminal device may be a desktop computer, laptop, handheld computer, or cloud server, etc. The terminal device may include, but is not limited to, a processor and a memory.

[0089] The processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor. The processor is the control center of the terminal device, connecting all parts of the terminal device via various interfaces and lines.

[0090] Based on the above-described method embodiments, another embodiment of the present invention provides a computer-readable storage medium including a stored computer program, wherein, when the computer program is executed, it controls the device where the computer-readable storage medium is located to execute a power system switching control method as described in any of the above-described method embodiments of the present invention.

[0091] The modules / units integrated in the device / terminal equipment, if implemented as software functional units and sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the above embodiments of the present invention can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying the computer program code, a recording medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium, etc.

[0092] Based on the above-described method embodiments, another embodiment of the present invention provides a computer program product, including a computer program or instructions, which, when executed by a communication device, implements a switching control method for a power system.

[0093] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A switching control method for a power system, characterized in that, Applied to an electric-hydrogen hybrid power system, the method includes: Collect the operating status data of the electric-hydrogen hybrid power system within a preset time period; The state change amount and state change rate are calculated using the operating state data. The operating state data, the state change amount, and the state change rate are combined to obtain a state feature sequence. The state feature sequence is input into a preset switching prediction model. The hidden layer in the switching prediction model is used to process the state feature sequence to obtain hidden state features. Each prediction branch in the switching prediction model is used to predict the hidden state features to obtain the switching probability, the switching behavior category, and the remaining switching time of the electric-hydrogen hybrid power system. If the switching probability, the switching behavior category, and the remaining switching time meet the preset switching rules, the switching behavior category is searched in a preset mode parameter mapping table to obtain the switching control parameters corresponding to the switching behavior category. Based on the remaining switching time, the control parameters to be switched, and the operating status data, a switching control strategy is generated to control the switching of the electric-hydrogen hybrid power system.

2. The switching control method for a power system as described in claim 1, characterized in that, The calculation of the state change amount and state change rate using the operating state data specifically includes: Several voltage changes are obtained by calculating several bus voltages of the DC bus in the operating status data. The operating status data includes several bus voltages of the DC bus, several bus currents, and several target power values ​​of the electric-hydrogen hybrid power system within the time period. By using the current of each of the busbars, several current changes of the DC busbars are obtained; The bus voltage and bus current of the DC bus at the same acquisition time are calculated to obtain several bus output powers of the DC bus within the time period, and based on the output power of each bus, several bus power changes of the DC bus are calculated. By using the target power values, several target power changes are obtained; The state change is obtained by summing up the voltage change, current change, bus power change and target power change. Based on the power change of each bus and the sampling period, the state change rate is calculated, wherein the sampling period is the time interval between two adjacent acquisitions in the operating state data.

3. The switching control method for a power system as described in claim 2, characterized in that, The process of summarizing the voltage changes, current changes, bus power changes, and target power changes to obtain the state change quantity specifically includes: The terminal voltage and charging / discharging current of the power battery at the same acquisition time in the operating status data are calculated to obtain several charging / discharging powers of the power battery within the time period. The electric-hydrogen hybrid power system also includes the power battery and the fuel cell. The operating status data includes several terminal voltages and several charging / discharging currents of the power battery, and several output voltages and several output currents of the fuel cell. Based on the charging and discharging power, several power power changes are calculated. The output voltage and output current of the fuel cell at the same acquisition time are calculated to obtain several fuel output powers of the fuel cell within the time period, and several fuel power changes are calculated based on each fuel output power. Using the charging and discharging power of the power battery, the fuel output power of the fuel cell, and the bus output power of the DC bus at the same acquisition time, several source-side power distribution ratios of the electric-hydrogen hybrid power system within the time period are calculated. Based on the power allocation ratios of each source side, several ratio changes are calculated; The state change is obtained by summing up each of the voltage changes, current changes, bus power changes, target power changes, power changes, fuel power changes, and proportional changes.

4. The switching control method for a power system as described in claim 3, characterized in that, The method of calculating several source-side power allocation ratios of the electric-hydrogen hybrid power system within the time period using the charging and discharging power of the power battery, the fuel output power of the fuel cell, and the bus output power of the DC bus at the same acquisition time, specifically includes: When the power battery is discharging, the charging and discharging power of the power battery and the fuel output power of the fuel cell at the same acquisition time are used to calculate several first power distribution ratios of the electric-hydrogen hybrid power system within the time period. When the power battery is charging, the fuel output power of the fuel cell and the bus output power of the DC bus at the same acquisition time are used to calculate several second power distribution ratios of the electric-hydrogen hybrid power system within the time period. By summing up the first power allocation ratios and the second power allocation ratios, a number of source-side power allocation ratios of the electric-hydrogen hybrid power system within the time period are obtained.

5. The switching control method for a power system as described in claim 1, characterized in that, If the switching probability, the type of behavior to be switched, and the remaining switching time satisfy a preset switching rule, then the type of behavior to be switched is retrieved from a preset mode parameter mapping table to obtain the switching control parameters corresponding to the type of behavior to be switched, specifically including: If the switching probability, the switching behavior category, and the remaining switching time satisfy the switching rules, then the switching behavior category is used to search in the mode parameter mapping table to obtain the fuel cell switching power, power battery switching power, and power change rate switching upper limit corresponding to the switching behavior category. The switching power of the fuel cell, the switching power of the power battery, and the upper limit of the power change rate are summarized to obtain the control parameters to be switched.

6. The switching control method for a power system as described in claim 1, characterized in that, The prediction branch includes a probability branch, a category branch, and a time branch. The hidden state features are predicted using each prediction branch of the switching prediction model to obtain the switching probability, the category of the behavior to be switched, and the remaining switching time of the electric-hydrogen hybrid power system. Specifically, this includes: The hidden state features are processed using the first fully connected layer in the probability branch to obtain a first intermediate quantity, and the first intermediate quantity is normalized to obtain the switching probability. The hidden state features are processed using the second fully connected layer in the category branch to obtain the second intermediate quantity; The second intermediate quantity is normalized to obtain the behavior probabilities corresponding to several candidate behavior categories, and the behavior probabilities are filtered to determine the behavior category to be switched. The hidden state features are processed using the third fully connected layer in the time branch to obtain the third intermediate quantity; The third intermediate quantity is processed using a preset linear rectification function to obtain the remaining switching time.

7. The switching control method for a power system as described in claim 1, characterized in that, The step of generating a switching control strategy based on the remaining switching time, the control parameters to be switched, and the operating status data specifically includes: Based on the remaining switching time, the switching transition time for the electric-hydrogen hybrid power system to switch from the current behavior to the behavior category to be switched is calculated; Using the switching transition time as the interpolation interval, interpolation processing is performed on the control parameters to be switched and the operating status data to obtain the switching parameters corresponding to each moment within the switching transition time. The switching parameters are summarized in chronological order to obtain the switching control strategy.

8. A switching control system for a power system, characterized in that, The switching control system is used to execute the switching control method of the power system as described in any one of claims 1 to 7.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a stored computer program, wherein, when the computer program is executed, it controls the device or apparatus containing the computer-readable storage medium to perform the power system switching control method as described in any one of claims 1 to 7.

10. A computer program product, comprising a computer program or instructions, characterized in that, When the computer program or instructions are executed by the communication device, the switching control method of the power system as described in any one of claims 1 to 7 is implemented.

Images