An intelligent charge-discharge management method and system for an aviation battery

Abstract

The application relates to the technical field of battery management, in particular to an intelligent charge-discharge management method and system for an aviation storage battery. The method comprises the following steps: superimposing a zero-mean symmetric current waveform on a pilot sub-string in a floating state, combining synchronous sampling of voltage and temperature, obtaining a voltage holding transient kernel and a thermal readiness transient kernel, thereby characterizing the potential holding capacity and temperature rise accessibility of the sub-string, finally predicting the voltage holding trajectory and temperature rise trajectory through kernel playback, and extracting key parameters as response indicators. The application can evaluate the health state and dynamic characteristics of the aviation storage battery in real time under the premise of not interrupting the safety of flight power supply, and provides support for high reliability management and life extension.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of battery management, in particular to an intelligent charge-discharge management method and system for an aviation storage battery. BACKGROUND

[0002] The existing aviation storage battery management method generally controls all battery modules as a unified object, lacks battery modules specially used for feature extraction and dynamic modeling, cannot accurately identify the voltage retention capability and temperature rise characteristics of the battery in the floating charge stage, and the accuracy of the health state evaluation is insufficient. At the same time, the existing technology usually adopts overall connection or simple parallel connection in the process of connecting multiple battery modules to the bus, and it is difficult to avoid the imbalance caused by current impact and voltage difference between sub-strings, which is easy to cause the instability of system operation and the attenuation of service life. In addition, the existing scheme is insufficient in utilizing bus voltage disturbance and micro-ripple signals, and cannot form a predictable voltage response and temperature response model, so it is difficult to meet the demand of fine management in the aviation application scene with high safety and high reliability requirements.

[0003] To solve the above problems, the present application designs an intelligent charge-discharge management method and system for an aviation storage battery. SUMMARY

[0004] The present application provides an intelligent charge-discharge management method and system for an aviation storage battery to solve the problems of the prior art. By superimposing a zero-mean symmetric current waveform on the pilot sub-string in the floating charge state, and combining the synchronous sampling of voltage and temperature, the voltage retention transient kernel and thermal readiness transient kernel are obtained, so as to characterize the potential retention capability and temperature rise accessibility of the sub-string. Finally, the voltage retention trajectory and temperature rise trajectory are predicted through kernel playback, and the key parameters are extracted as response indicators.

[0005] To achieve the above purpose, the present application provides the following technical scheme:

[0006] An intelligent charge-discharge management method for an aviation storage battery is applied to a flight control battery, the flight control battery includes a bus connected with a flight control system and a plurality of power supply modules, and the method comprises:

[0007] By the control of the battery management system, at least one power supply module in the flight control battery is determined as a pilot sub-string, and the remaining power supply modules are determined as main sub-strings;

[0008] When the flight control battery is in a floating charge state, a zero-mean symmetric current waveform is superimposed on the pilot sub-string in the floating charge current, and the voltage response and temperature response are calculated;

[0009] determining a dynamic internal resistance and a polarization degree of the pilot sub-string according to the voltage response and the temperature response, and adjusting the temperature of the pilot sub-string to a preset micro-thermal window and adjusting the voltage of the pilot sub-string to a voltage range of the bus through closed-loop control;

[0010] when the main power supply fails, connecting the pilot sub-string to the bus according to a switching criterion to maintain the bus voltage, and grouping the main sub-strings and connecting them to the bus through a soft start strategy for limiting the current change rate during the process of connecting the main sub-strings to the bus.

[0011] determining at least one power supply module in the flight control battery as a pilot sub-string, comprising:

[0012] in a state where each power supply module is electrically isolated from the bus, establishing a voltage following closed loop of the terminal voltage to the bus reference voltage through a bidirectional DC converter connected with the corresponding power supply module, wherein the voltage following closed loop takes the duty cycle of the bidirectional DC converter as a control quantity;

[0013] according to the voltage following closed loop, calculating the voltage tracking error and the control quantity change of each power supply module in the case of converting to a unified reference waveform;

[0014] determining the power supply module with a tracking error less than or equal to a first threshold value and a control quantity change less than or equal to a second threshold value within a preset time window as a pilot sub-string.

[0015] for the pilot sub-string, superimposing a zero-mean symmetric current waveform in the floating current, and calculating a voltage response and a temperature response, comprising:

[0016] obtaining bus micro-ripple data recorded in historical flight tasks;

[0017] performing principal component decomposition on the bus micro-ripple data, selecting the first K spectral bases to obtain a symmetric sequence, and processing the symmetric sequence through an optimization algorithm under the condition of meeting a preset constraint condition to obtain two symmetric current sequences that are mutually orthogonal and have zero net electric quantity, wherein the constraint condition includes electromagnetic compatibility constraint and micro-thermal energy consumption constraint, and the optimization objective function of the optimization algorithm is calculated through the bus micro-ripple data recorded in the historical flight tasks;

[0018] superimposing the symmetric current sequence to the floating current of the pilot sub-string, and obtaining an original sequence of terminal voltage and an original sequence of temperature of the pilot sub-string according to a sampling time sequence synchronized with the symmetric current sequence;

[0019] performing orthogonal demultiplexing and matched filter array projection on the original sequence of terminal voltage and the original sequence of temperature to obtain a voltage characteristic coefficient vector and a temperature characteristic coefficient vector corresponding to the spectral bases;

[0020] processing the voltage feature coefficient vector and the temperature feature coefficient vector through compressive sensing to obtain a voltage retention transient kernel and a thermal readiness transient kernel, wherein the voltage retention transient kernel represents a potential retention capability of the pilot substring, and the thermal readiness transient kernel represents a temperature rise accessibility of the pilot substring;

[0021] nuclear playback of a standard excitation template corresponding to the symmetric current sequence according to the voltage retention transient kernel and the thermal readiness transient kernel to obtain a voltage retention prediction trajectory and a temperature rise prediction trajectory, and determination of a maximum voltage offset, a recovery slope, and a retention time length in the voltage retention prediction trajectory as a voltage response, and determination of a time to a micro-thermal window, a steady-state temperature difference, and an overshoot amount in the temperature rise prediction trajectory as a temperature response.

[0022] The symmetric sequence is processed through an optimization algorithm under the condition that the preset constraint condition is met to obtain two symmetric current sequences that are mutually orthogonal and have zero net electric quantity, including:

[0023] The symmetric sequence is decomposed according to frequency components and time domain components, and the decomposed vectors are weighted through a preset weight to obtain a candidate sequence set, wherein the weight is calculated according to electromagnetic compatibility constraints and micro-thermal energy consumption constraints;

[0024] The candidate sequence set is input into a preset optimization algorithm to calculate a net electric quantity deviation, a frequency spectrum distribution, and an energy consumption of the candidate sequence to obtain a first current sequence and a second current sequence;

[0025] The first current sequence and the second current sequence are modified through amplitude adjustment and phase rotation to make the first current sequence and the second current sequence mutually orthogonal within a preset time window.

[0026] Orthogonal de-multiplexing and matched filtering array projection are performed on the terminal voltage original sequence and the temperature original sequence, including:

[0027] The terminal voltage original sequence and the temperature original sequence are decomposed in a feature space constituted by the spectral bases to obtain a plurality of independent response components;

[0028] A projection matrix is calculated through bus micro-ripple data recorded in historical flight missions, and the current bus micro-ripple data is used as a projection operator;

[0029] The response components corresponding to the terminal voltage original sequence and the temperature original sequence are mapped into voltage feature coefficient vectors and temperature feature coefficient vectors corresponding to the spectral bases in combination with the projection matrix and the projection operator.

[0030] The processing of the voltage feature coefficient vector and the temperature feature coefficient vector through compressive sensing includes:

[0031] constructing a measurement matrix according to the bus micro ripple data in historical flight missions, and projecting the voltage feature coefficient vector and the temperature feature coefficient vector to an observation domain corresponding to the measurement matrix;

[0032] sparsely representing the projected observation vector under a preset sparse dictionary, wherein the sparse dictionary is calculated through a spectral basis;

[0033] According to the sparse representation, solving an optimal sparse coefficient vector through a norm-constrained sparse recovery algorithm to reconstruct dynamic response signals corresponding to the voltage and the temperature;

[0034] extracting transient features of the dynamic response signals to obtain a voltage retention transient kernel and a thermal readiness transient kernel.

[0035] The determination of the dynamic internal resistance and the polarization degree of the navigation substring includes:

[0036] dividing the maximum voltage offset in the voltage retention prediction trajectory by the effective amplitude of the symmetric current sequence to obtain the dynamic internal resistance;

[0037] fitting the recovery slope in the voltage retention prediction trajectory with the retention duration to obtain a polarization time constant;

[0038] mapping the polarization time constant with a steady-state temperature difference to determine an initial polarization degree, and correcting the initial polarization degree in combination with an overshoot amount in the temperature rise prediction trajectory to obtain the polarization degree.

[0039] The switching criterion includes satisfying a zero voltage or zero current condition when the navigation substring is connected to the bus.

[0040] The grouping and connection of the main substrings to the bus through the soft start strategy include:

[0041] grouping the main substrings according to voltage change rates, and performing bus voltage alignment and current-limiting pre-charging before each group is connected;

[0042] During the connection process, the current change rate is limited by adjusting the duty cycle of the bidirectional DC converter, and the voltage, temperature and current stability of the main substrings in the group are monitored;

[0043] When the current group reaches a preset steady-state condition, the next group is connected, and all the main substrings are connected to the bus.

[0044] An intelligent charge-discharge management system for an aviation storage battery, the system comprising:

[0045] a battery management module for determining navigation substrings and main substrings, and executing a charge-discharge control strategy;

[0046] a signal modulation module, configured to superimpose a symmetric current sequence in the floating current, and collect voltage response and temperature response of the pilot substring;

[0047] a coordination control module, configured to determine dynamic internal resistance and polarization degree according to the voltage response and the temperature response, and access the bus according to switching criteria and soft start strategy when the main power fails.

[0048] Compared with the prior art, the application has the following beneficial effects:

[0049] By setting the pilot substring, the application introduces a zero-mean symmetric current sequence in the floating state and extracts the voltage retention transient kernel and the thermal readiness transient kernel, realizes real-time modeling and prediction of the battery dynamic characteristics, and can accurately evaluate the voltage response and the temperature response without affecting the energy supply of the pilot substring. Compared with the prior art, the application avoids the defect that the overall substring is difficult to directly obtain dynamic characteristics, realizes more refined energy management in the aviation battery application scenario, prolongs the service life of the battery, and improves the reliability and safety of the system. BRIEF DESCRIPTION OF DRAWINGS

[0050] Other features, objects and advantages of the application will become more apparent after reading the following detailed description of non-limiting embodiments, made with reference to the accompanying drawings:

[0051] Figure 1 A problem principle schematic diagram is provided for the embodiments of the application;

[0052] Figure 2 An exemplary application scenario diagram is provided for the embodiments of the application;

[0053] Figure 3 A flowchart of an intelligent charge-discharge management method for an aviation battery is provided for the embodiments of the application;

[0054] Figure 4 A pilot substring response calculation flowchart is provided for the embodiments of the application. DETAILED DESCRIPTION

[0055] The technical solutions in the embodiments of the application will be described clearly and completely below with reference to the accompanying drawings in the embodiments of the application. Obviously, the described embodiments are only part of the embodiments of the application, not all embodiments of the application.

[0056] Reference herein to an "embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment can be included in at least one embodiment of the application. The appearances of the phrase in various places in the specification are not necessarily all referring to the same embodiment, nor are they necessarily mutually exclusive of one another. Those skilled in the art will recognize that the embodiments described herein can be combined with other embodiments.

[0057] The application is suitable for the battery pack composed of multiple modules in parallel in the aircraft flight control power supply system. The charge and discharge management process is limited by the stability of the power bus and the energy supply redundancy under the flight task. The traditional management mode is mostly based on fixed balance or static switching, which is difficult to balance dynamic performance and safety margin. In the power supply process, limited by the fluctuation of external main power and the transient load characteristics of airborne equipment, voltage inconsistency, polarization difference and uneven heat distribution may occur between battery modules, which may further cause current impact and bus voltage disturbance when switching or merging power supply.

[0058] Application scenarios include but are not limited to:

[0059] Emergency power starting battery, which needs to take over and maintain the bus voltage quickly in the case of main power failure;

[0060] Modular battery system with multiple sub-strings in parallel, which has internal differences and uneven dynamic response problems;

[0061] Aircraft battery under long-time float charging condition, polarization characteristics and internal resistance changes have an impact on power supply safety.

[0062] The selection of application scenarios is based on the common characteristics of aviation batteries, which at least include one of the following:

[0063] There is dynamic inconsistency in the module parallel system, and a reliable navigation sub-string needs to be determined to realize the transition;

[0064] It is difficult to identify the internal parameters of the sub-string in the float state in real time, and the voltage and temperature response needs to be extracted through perturbation means;

[0065] Smooth switching and soft start are required when the main power fails to avoid transient impact on flight control safety.

[0066] It should be noted that the intelligent charge and discharge management method proposed in the application does not depend on the fixed configuration or single circuit structure of the battery module, but is aimed at the aviation power supply scene with multiple modules in parallel, uneven response and external power constraints.

[0067] It is worth mentioning that the strategy described in this application is not only applicable to the flight control battery, but also can be applied to other systems that require high reliability power switching, such as spacecraft energy management, unmanned aerial vehicle power backup or high safety level ground power system.

[0068] It can be understood that in this application, the flight control battery specifically refers to the emergency starting battery of the aircraft. In the normal flight state, the bus is maintained stable by the main generator or external ground power supply, and the flight control battery is in long-term float charging state, and its main function is to keep full power for emergency use. Due to the extremely high safety level of the aviation power supply system, the flight control battery not only has the function of conventional energy storage, but also must have the functions of emergency starting and bus maintenance in emergency situations.

[0069] The emergency starting function is that when the main power system of the aircraft such as the engine or auxiliary power device needs to be restarted on the ground or in the air, the flight control battery releases energy in the form of large current to provide instantaneous power for the starting motor or control unit. This requires low internal impedance and fast polarization response of the battery, otherwise it will cause the starting to fail.

[0070] The bus maintenance function is that when the main generator fails, the flight control battery must take over the bus voltage instantaneously to smooth the power supply and avoid the failure of the flight control computer or critical actuator. At this time, not only the voltage stability needs to be ensured, but also the impact current or thermal imbalance caused by the difference in internal resistance between the parallel modules needs to be avoided.

[0071] In some optional embodiments, the intelligent charge and discharge management method of the flight control battery can be realized in combination with the existing wiring structure and capacity recovery method of the flight control battery panel. By introducing a zero-mean symmetric current waveform in the float state, it can be combined with the standard constant current discharge condition to form a dynamic response incentive for the pilot sub-string. On this basis, using the capacity calculation formula, the capacity recovery trend and voltage retention characteristics of the module under disturbance incentive can be evaluated in real time without damaging the overall battery performance, thereby providing experimental conditions for extracting dynamic internal resistance and polarization degree. This way avoids the damage to the battery caused by traditional over-discharge or over-charge detection, and is more in line with the strict safety requirements of aviation applications.

[0072] In other optional embodiments, the pilot sub-string and the main sub-string can be connected to the bus through different test interfaces, and the TRU-TEST or FCC-BAT-TEST signal can be used to introduce and feedback the voltage disturbance. This way can realize dynamic monitoring and comparison of bus voltage without affecting normal power supply, so as to obtain higher resolution voltage response and temperature response data. Unlike traditional single-point open-circuit voltage screening, this method relies on the existing multi-point signal interface of the flight control panel, making the pilot sub-string more engineering operable in the selection and verification link.

[0073] In one example, the requirement that the discharge termination voltage be maintained above 21.6V, and the activation and recovery through the 28V bus power combined with 2 / 1, 2 / 7 pin signals, can be combined with a zero-mean symmetric current superposition, i.e., a small perturbation signal is added in the capacity recovery and activation link, for detecting different module voltage retention transient kernels and thermal readiness transient kernels. This can complete the modeling of the dynamic characteristics of the module in the conventional maintenance and detection process, reduce additional test links, and improve detection efficiency.

[0074] It is easy to understand that the navigation substring dynamic response calculation method proposed in the present application can not only be compatible with existing flight control battery detection processes, such as constant current source, 28V bus power supply and pin signal control, but also can realize synchronous analysis of the voltage and temperature double channels through superposition of symmetric current in the floating state. This not only retains the safety boundary of the traditional method, but also increases the discriminant ability in the dynamic dimension, provides a more accurate prerequisite for the soft start of the subsequent main substring, and improves the reliability and safety of the flight control battery as a whole.

[0075] It can be understood that since the flight control battery is in a floating state for a long time, the battery cell will have problems such as decrease in electrochemical activity, increase in polarization difference and uneven heat distribution. Simple voltage balancing or static health monitoring often cannot reflect the dynamic response characteristics. Therefore, in the present application, a zero-mean symmetric current sequence is superimposed to stimulate the transient response of the battery, and the voltage retention capability and thermal readiness capability are extracted through de-multiplexing and compressed sensing. It can dynamically depict whether the battery still has the ability to quickly start and stably take over.

[0076] Reference Figure 1 , Figure 1 The problem principle schematic diagram for the embodiments of the present application is provided.

[0077] Figure 1 Taking a space shuttle as an example, the flight control battery carried by the space shuttle is an emergency starting power and key power support unit, which provides energy to various electrical equipment on board such as flight control systems, communication equipment, navigation instruments, etc. through the bus. The flight control battery needs to maintain stable output in various operating states to ensure flight safety.

[0078] Figure 1 It is shown that when the main power source fails in the flight task, i.e., at T1 shown in FIG. 1, Figure 1 It is shown that the flight control battery needs to immediately take over the power supply to maintain the normal operation of the flight control system and key equipment, however, Figure 1 It is further shown that typical problems that can occur in the takeover process:

[0079] Due to the sudden disconnection of the main power supply, the voltage of the flight control battery may drop instantaneously at the beginning of the takeover, and may be accompanied by oscillatory fluctuations during the recovery process, affecting the stability of the flight control system.

[0080] In the case of continuous high-current discharge or frequent sudden power supply switching, the temperature of the flight control battery rises rapidly, and if there is no effective management and protection mechanism, it is easy to cause thermal runaway, endangering the safety of the entire flight mission.

[0081] In an embodiment not shown in the figure, when the sudden load or critical equipment is started, the battery output current increases sharply, causing the bus voltage to fluctuate again, and even causing the power equipment to malfunction or drop.

[0082] Reference Figure 2 , Figure 2 An example application scenario is provided for the embodiments of the present application.

[0083] Figure 2 The power management scenario of the aircraft during the flight mission is shown: the flight control battery and the main power supply are connected to the bus to provide energy for the onboard power equipment. In normal state, the main power supply undertakes the main power supply task, and the flight control battery is in standby or charging state; when the main power supply drops or power is insufficient during flight, the flight control battery switches to discharge state to take over the bus power supply to ensure the continuous and stable operation of the critical power equipment.

[0084] Figure 2 Further, the flight control battery is divided into a navigation sub-string and a main sub-string. The navigation sub-string can be preferentially put into work at the moment of main power failure or switching, quickly establishing a stable voltage platform to provide seamless protection for the flight control system and critical power equipment. The main sub-string is connected after completing the short-time voltage transition to undertake subsequent continuous energy supply. The structure effectively avoids the voltage drop and delay problem of a single battery pack during emergency switching, improving the overall safety and stability of the system.

[0085] It can be understood that the specific number of the navigation sub-string and the main sub-string is not limited by the present application, and its configuration mode can be flexibly adjusted according to the power demand, power supply architecture and power supply redundancy design of different types of aircraft. For example, the navigation sub-string can be composed of several series of battery cells, which is preferentially used to undertake the functions of rapid response and bus voltage stabilization; the main sub-string provides main energy support for long-time and high-power continuous power supply.

[0086] Next, combined with the accompanying Figure 3 , an intelligent charge and discharge management method for an aviation battery is introduced. The present application is applied to a flight control battery, which includes a bus connected to a flight control system and a plurality of power supply modules. The method comprises:

[0087] S1: Determine at least one power supply module in the flight control battery as a pilot substring through battery management system control, and determine the remaining power supply modules as main substrings;

[0088] In this embodiment, the pilot substring is responsible for voltage establishment and dynamic response, and the main substring is mainly used for energy support. This division enables the pilot substring to quickly take over the bus voltage in an emergency state, avoiding voltage drop during switching, while ensuring that the main substring has sufficient capacity for long-term power supply. This solves the problem of slow response and unstable voltage during emergency switching of traditional single battery groups, and improves the safety guarantee capability of the flight control battery during critical tasks.

[0089] S2: When the flight control battery is in a floating state, for the pilot substring, superimpose a zero-mean symmetric current waveform on the floating current, and calculate the voltage response and temperature response;

[0090] In this embodiment, by superimposing a zero-mean symmetric current waveform, the characteristics of the battery can be dynamically stimulated without changing the overall energy state. By collecting voltage and temperature responses, data reflecting the health and activation level of the battery can be obtained. Compared with conventional static detection, it is more precise and can be used for online diagnosis during non-task flight, thereby identifying potential aging or abnormal conditions in advance and reducing the risk of battery failure during task execution.

[0091] S3: Determine the dynamic internal resistance and polarization degree of the pilot substring according to the voltage response and temperature response, and adjust the temperature of the pilot substring to a preset micro-heat window and the voltage of the pilot substring to the voltage range of the bus through closed-loop control;

[0092] In this embodiment, the pilot substring works in a stable micro-heat window through closed-loop control, thereby maintaining the activity of electrochemical reactions and delaying performance degradation. At the same time, the voltage is automatically adjusted to the bus range to ensure seamless access when switching occurs, avoiding additional voltage fluctuations. This effectively solves the problem of inaccurate single-module parameter control and inability to ensure availability in existing battery management.

[0093] S4: When the main power supply fails, connect the pilot substring to the bus according to the switching criteria to maintain the bus voltage, and group the main substring and connect it to the bus through a soft start strategy;

[0094] In this embodiment, the pilot substring first independently undertakes the bus power supply task, keeping the flight control device stable operation; then the main substring is gradually put into the group in a grouped manner, and the soft start strategy can avoid the risk of instantaneous large current impact, while improving the voltage stability. In this way, it can not only ensure the continuous power supply of key equipment in extreme cases, but also prolong the overall life of the battery.

[0095] Those skilled in the art can understand that the switching criterion can be set according to the bus voltage threshold, current state and flight task demand, which is not limited in this application.

[0096] Before the specific technical content corresponding to the unfolding step, the embodiment of the present application needs to be emphasized again.

[0097] In a typical aviation flight control power supply configuration, the flight control battery often needs to consider both fast response and long-term stable power supply. However, due to the problems of cell consistency, load fluctuation and sudden failure, etc., the traditional battery pack often appears response lag, voltage drop and parallel impact, etc., which is difficult to guarantee the continuous support of the bus voltage at the critical moment. The scheme adopted in this embodiment is not simply from the single battery or charging method, but through the combination of overall topology and dynamic management logic, a pilot substring and main substring architecture with division of labor and cooperation relationship is established.

[0098] Under the pilot substring and main substring architecture, at least one power supply module is preset as the pilot substring, which can obtain dynamic parameters through disturbance diagnosis at ordinary times. When in the floating state, a symmetric current waveform with zero mean value is superimposed on the pilot substring, so that the internal resistance change and polarization trend are exposed without changing the overall state of charge, thereby obtaining more accurate indicators than static sampling.

[0099] Those skilled in the art can understand that the form of data collection is not limited, which can be voltage, current, temperature and other common signals, or other data that can reflect the dynamic response characteristics, as long as it meets the minimum requirement of identifying the operating state. In this way, the pilot substring not only assumes the role of conventional energy, but also becomes a reference object for online monitoring.

[0100] Further, according to the temperature response and voltage response of the pilot substring, the application proposes a micro-heat window adjustment strategy. By closed-loop control, the temperature of the pilot substring is maintained within a preset narrow range, which can stimulate the activity of electrochemical reaction without increasing the overall thermal risk, so that the cell can maintain a low polarization level in long-term work. At the same time, the voltage is constrained within the bus range, which creates conditions for subsequent cut-in operation. This pre-activation mechanism that is always available effectively avoids the bus oscillation caused by state mismatch when suddenly accessing.

[0101] Further, in the case of failure of the main power supply, the navigation sub-string is first connected to the bus, and by virtue of its pre-adjusted voltage and temperature conditions, it can quickly stabilize the bus voltage. Subsequently, the main sub-string is not connected in parallel at one time, but is gradually connected in groups in a soft start manner, which not only reduces the hardware stress caused by large current impact, but also reduces the influence of voltage disturbance on the flight control device. This process not only guarantees short-time response, but also takes into account long-term power supply capability, significantly improving the safety and reliability of the flight control battery in complex flight tasks.

[0102] Next, the technical content of the power supply module division of the embodiments of the present application is further expanded.

[0103] It can be understood that the power supply module is not randomly divided into navigation sub-strings and main sub-strings, but is optimized based on the matching degree with the bus voltage and its own characteristics. The bus, in the aviation electrical system, generally refers to a public electrical channel that plays a unified current collection and distribution role, and its voltage level directly determines the power supply stability of the flight control devices, navigation instruments and communication units connected thereto. Therefore, the bus voltage is not only the reference point of the external load, but also an important reference condition for the selection of the power supply module.

[0104] In specific applications, the module with a smaller voltage level deviation from the bus voltage is preferentially selected as the navigation sub-string, so that it can quickly enter the available state without the need for additional substantial adjustment. In this way, not only the energy consumption and time consumption of subsequent closed-loop regulation are reduced, but also the takeover is more efficiently completed in the case of failure of the main power supply. At the same time, the remaining modules are determined as the main sub-string, which is gradually connected to the bus through grouping and soft start to realize the expansion of the overall capacity. This division method can significantly reduce the fluctuation amplitude of the bus voltage and avoid transient impact or power interruption caused by excessive differences in module parameters.

[0105] In some optional embodiments, the determination criteria are not limited to voltage matching, but can also consider factors such as internal resistance size, temperature rise characteristics and historical operation data, and the weights can be flexibly set according to actual operation requirements. As long as the requirements for rapid support of the bus voltage and overall power supply stability are met, the present application does not make more limitations.

[0106] In one example, determining at least one power supply module in the flight control battery as a navigation sub-string includes:

[0107] S1.1: In the state of electrical isolation of each power supply module from the bus, a voltage following closed loop of an end voltage to a bus reference voltage is established through a bidirectional DC converter connected with the corresponding power supply module, wherein the voltage following closed loop takes the duty cycle of the bidirectional DC converter as the control quantity;

[0108] Specifically, in the case where each power supply module has not been directly connected in parallel with the bus, the problem of excessive instantaneous voltage difference caused by direct connection may occur, which may cause impact current and cause damage to the module itself or a large fluctuation in the bus voltage.

[0109] In the present embodiment, each power supply module establishes a voltage following closed loop of voltage to bus reference voltage in an electrically isolated state through a corresponding bidirectional DC converter. The bidirectional DC converter can be flexibly switched between boost and buck modes, and the control quantity is the duty cycle change of the internal power switch tube. By adjusting the duty cycle, the module port voltage can be changed in real time to make it as close as possible to the reference voltage of the bus. In this way, even if there is a certain difference in the open circuit voltage of the module itself, it can be compensated by dynamic adjustment of the closed loop.

[0110] It can be understood that through the foregoing setting, not only the controllable following of the voltage of each module port is realized, but also a detection basis for subsequently selecting a module with optimal voltage stability is determined, while the uncertainty risk caused by directly connecting the module to the bus without control is avoided.

[0111] Further, this following closed loop can respond to the fluctuation of the bus reference voltage in real time to ensure that the test environment is consistent with the actual connection working condition, thereby improving the reliability of the judgment.

[0112] S1.2: According to the voltage following closed loop, calculate the voltage tracking error and control quantity change of each power supply module under the condition of transforming into a unified reference waveform;

[0113] In the present embodiment, the port voltage signals of each module under closed loop control are uniformly transformed into the same reference waveform to eliminate the interference caused by the change of the bus reference waveform, thereby ensuring the consistency of the comparison reference.

[0114] Further, under this condition, the voltage tracking error and control quantity change of each power supply module are calculated respectively. The voltage tracking error reflects the degree of approximation of the module voltage to the bus reference voltage, and the smaller the value, the higher the matching between the module and the bus. The control quantity change represents the fluctuation amplitude of the duty cycle of the bidirectional DC converter, and its size directly reflects the adjustment strength required by the module to maintain stable output. If the duty cycle of a certain module needs to be frequently and substantially adjusted in the process of realizing voltage following, it indicates that there is inherent difference or fluctuation source between its own voltage and the bus reference voltage, and its long-term connection to the bus will increase the system adjustment burden.

[0115] It can be understood that the combination of the two types of parameters can comprehensively evaluate the controllability and stability of the module. The processing of the unified reference waveform eliminates the influence of the bus instantaneous disturbance, so that the measured tracking error and the control amount change truly reflect the performance of the module itself rather than external environmental factors, thereby improving the accuracy of the determination.

[0116] S1.3: determining the power supply module with the tracking error less than or equal to the first threshold value and the control amount change less than or equal to the second threshold value in the preset time window as the leading sub-string;

[0117] In the embodiment, a preset time window is set for observing the performance of each module in the continuous working process. If the voltage tracking error of a certain module is always less than or equal to the first threshold value and the corresponding control amount change is less than or equal to the second threshold value in the preset time window, it can be considered that the module can stably follow the bus reference voltage and has a smaller adjustment burden. The module meeting the above double conditions is determined as the leading sub-string.

[0118] It can be understood that this determination method avoids the errors that may be caused by the instantaneous state alone, and ensures that the selected module has stability and consistency in actual operation.

[0119] Further, the threshold values can be set flexibly in combination with the characteristics of different types of modules and the actual bus voltage fluctuation tolerance, thereby realizing adaptability under the premise of ensuring reliability.

[0120] It is easy to understand that the deviation between the terminal voltage of the leading sub-string selected by this method and the bus reference voltage is relatively small when the leading sub-string is connected to the bus, so the required dynamic adjustment amount is significantly reduced, and the preliminary support for the bus can be completed in a short time. This way ensures that the bus voltage does not fluctuate greatly in the event of a main power failure or system switching, thereby creating a stable environment for the gradual connection of other power supply modules.

[0121] It should be noted that the system has selected the module that is closest to the bus voltage range and has a more stable thermal response as the leading sub-string when detecting the dynamic characteristics of each power supply module. As a result, when the leading sub-string is connected to the bus, it is already in a small difference range with the bus voltage and has a suitable temperature state. After the leading sub-string is connected to the bus, the battery management system needs to make a small steady-state adjustment based on the electrical inertia and thermal inertia, such as relieving the voltage disturbance that may occur at the moment of connection through a smoothing control strategy, or maintaining a micro-thermal window through passive thermal management.

[0122] Reference Figure 4 , Figure 4 The flowchart of the leading sub-string response calculation provided by the embodiment of the application is shown.

[0123] Next, the technical content of the application embodiment about the response calculation of the navigation substring is further expanded.

[0124] It can be understood that the response calculation at this time occurs when the flight control battery is in a floating state. The so-called floating state can be understood as that the battery module is electrically connected with the bus, but does not undertake actual energy output task, and the charging current is only at a very low level, which is used to maintain the battery terminal voltage to be stable in the interval close to the bus reference voltage. In other words, in the floating state, the battery neither appears significant discharge power supply nor enters a substantial charging process, but is maintained in a low-power balance state to obtain stable external observation conditions without changing the SOC.

[0125] Further, simply relying on the open circuit voltage or static impedance to determine whether a module is suitable as a navigation substring often has deviation, because the dynamic disturbance faced by the battery module in the real running process will cause differences between the performance and the static index. If the dynamic characteristic verification is not performed before being connected to the bus, the navigation substring may have deficiencies in initial voltage maintenance and temperature rise control, thereby causing the bus voltage to fluctuate for a short time, affecting the overall system stability. Therefore, the application proposes to simulate the disturbance effect caused by the bus micro ripple in the flight task by artificially superimposing a zero-mean symmetric current waveform in the floating state, to reproduce the real performance of the battery under dynamic excitation in a controlled and repeatable manner.

[0126] It can be further understood that the zero-mean symmetric current does not have cumulative effect on the battery SOC, because the positive and negative half cycle energies cancel each other out, and only exert disturbance on the module terminal voltage and internal temperature distribution at the transient level. In this way, the system can observe the response speed and stability of the module to the external dynamic load without changing the battery charge level. For example, the voltage maintenance transient core reflects whether the potential of the module can quickly return to the reference value after being disturbed by a small disturbance, and the thermal readiness transient core reflects the heating rate and heat dissipation capacity of the module under continuous slight excitation.

[0127] In other words, the reason for calculating the voltage response and temperature response in the floating state is that the floating provides a stable observation window that does not interfere with the SOC, and the symmetric current excitation provides a test signal that simulates the actual disturbance. The combination of the two enables the system to accurately extract the dynamic performance index of the module without increasing energy consumption and damaging the battery life. Ultimately, it is these response characteristics that ensure that the selected navigation substring can enter a stable working state more quickly when it is actually connected to the bus, reducing the bus voltage regulation pressure.

[0128] In one example, for the navigation substring, a zero-mean symmetric current waveform is superimposed in the floating current, and the voltage response and temperature response are calculated, including:

[0129] S3.1: Obtain bus micro ripple data recorded in historical flight tasks;

[0130] Specifically, during multiple execution of tasks, the bus voltage of the aircraft is not always completely stable DC, but is superimposed with small periodic disturbances introduced by various electrical load switching, power converter modulation and environmental factors coupling. These disturbances are micro ripple signals in the time domain. If only idealized waveforms are constructed in the laboratory to simulate, it is difficult to truly reflect the actual dynamic characteristics of the bus under flight conditions.

[0131] In this embodiment, the bus voltage is first sampled with high precision by the battery management system in each flight task, and the recorded data is stored to obtain the original micro ripple data reflecting the dynamic characteristics of the bus under actual flight conditions.

[0132] It can be understood that the calling of real historical task data can ensure that the generated symmetric current sequence is more consistent with the electromagnetic interference spectrum of the flight environment, thereby achieving higher representativeness and reliability in the response calculation link of the pilot substring.

[0133] Further, in order to ensure data availability, the bus voltage signal is also band-limited sampled and anti-aliasing filtered during the collection process to ensure that the feature distribution of the micro ripple data in the frequency domain is not disturbed by noise and artifacts.

[0134] S3.2: Perform principal component decomposition on the bus micro ripple data, select the first K spectral bases to obtain a symmetric sequence, and process the symmetric sequence by an optimization algorithm under the condition that a preset constraint condition is met, to obtain two symmetric current sequences that are mutually orthogonal and have zero net electric quantity, wherein the constraint condition includes electromagnetic compatibility constraint and micro thermal energy consumption constraint, and an optimization objective function of the optimization algorithm is calculated by the bus micro ripple data recorded in historical flight tasks;

[0135] Specifically, after obtaining the bus micro ripple data, mathematical bases that can represent the main disturbance characteristics need to be extracted therefrom in order to be used to construct test current sequences. If the original waveform is directly superimposed on the floating current, it is difficult to ensure controllable disturbance energy, and non-target frequency band interference signals may be introduced, resulting in inability to accurately measure the response characteristics of the pilot substring.

[0136] In this embodiment, the principal component decomposition method is used to process the historical micro ripple data. By calculating the covariance matrix and performing eigenvalue decomposition, the first K spectral bases that can reflect the main energy distribution are extracted. On the one hand, data dimensionality reduction and redundancy removal can be achieved, and on the other hand, the main dynamics components of the bus ripple can be ensured to be reflected in the generated test sequence set, thereby improving the effectiveness and representativeness of the current superposition.

[0137] Further, after obtaining the spectrum base, the test current waveform is generated by constructing a symmetric sequence. The so-called symmetric sequence refers to strict symmetry between positive and negative half cycles in a complete cycle, so that the integral of the overall current waveform is zero, that is, the zero net charge characteristic, which can realize dynamic excitation without changing the state of charge of the battery, thereby avoiding SOC deviation and additional energy consumption. At the same time, the use of two mutually orthogonal symmetric current sequences can ensure comprehensive excitation and decoupling measurement of the navigation substring in different characteristic dimensions, preventing response overlap caused by single excitation.

[0138] Further, in the optimization process, the embodiment introduces two types of conditions, electromagnetic compatibility constraints and micro-thermal energy consumption constraints. The electromagnetic compatibility constraints are used to limit the generated current sequence to not exceed the established interference tolerance in the frequency spectrum, avoiding additional impact on the flight control system and communication equipment during the test process; the micro-thermal energy consumption constraints are used to limit the amplitude and duration of the test current, so that it does not cause the module to overheat or accelerate aging when applying disturbance. In order to meet these constraint conditions, the embodiment uses an optimization algorithm based on the objective function to solve, which is calculated from the energy distribution of the bus micro ripple in the historical flight mission. The optimization process is to balance the system compatibility and battery safety while maintaining true representation.

[0139] In one example, the specific steps of S3.2 are as follows:

[0140] S3.2.1: Decompose the symmetric sequence according to the frequency component and the time domain component, and weight the decomposed vectors by a pre-set weight to obtain a candidate sequence set, wherein the weight is calculated according to the electromagnetic compatibility constraints and the micro-thermal energy consumption constraints;

[0141] Specifically, the bus micro ripple usually contains components of multiple different frequency bands, and different frequency bands have differences in electromagnetic interference, thermal effects and dynamic response characteristics on the battery module. If the undecomposed symmetric sequence is used directly, it will cause excessive concentration of some high-frequency components in the sequence, thereby causing electromagnetic compatibility risks, or excessive low-frequency components, causing excessive heat accumulation in the substring.

[0142] In the embodiment, first, the original symmetric sequence is subjected to fast Fourier transform to map the time domain waveform to the frequency domain to obtain each frequency component contained in the signal. Then, according to a pre-set frequency band division interval, the frequency spectrum is divided into several sections, for example, the low-frequency section corresponds to the battery electrochemical polarization characteristics, the medium-frequency section corresponds to the bus parasitic impedance coupling characteristics, and the high-frequency section corresponds to the electromagnetic compatibility sensitive section. After preliminary division in the frequency domain, the components of each frequency section are mapped back to the time domain through inverse transformation, thereby obtaining a plurality of vectors with independent time domain characteristics, each of which corresponds to a typical disturbance mode.

[0143] Further, after the decomposition is completed, each vector needs to be weighted to form a candidate sequence set. The calculation of the weight is not a fixed value, but is dynamically determined according to the electromagnetic compatibility constraint and the micro-thermal energy consumption constraint. For example, when a certain frequency band falls near the sensitive frequency of the flight control device, the weight of the vector will be significantly reduced to avoid the generated test current from being too strong in the energy of the frequency band and interfering with the flight control communication link. Conversely, when a certain frequency band belongs to a battery thermal consumption sensitive section, the corresponding weight is set to a lower value by calculating the contribution of the battery temperature rise caused by the injection current of the frequency band, so as to ensure that the energy distribution does not cause local overheating.

[0144] In specific implementation, the electromagnetic compatibility spectrum constraint curve and the cell thermal resistance model can be pre-established, the two are superimposed to form a constraint function, and then the constraint function is used to assign a normalized weight factor to each vector.

[0145] S3.2.2: input the candidate sequence set into a preset optimization algorithm to calculate the net electric quantity deviation, the spectrum distribution and the energy consumption of the candidate sequence, and obtain the first current sequence and the second current sequence;

[0146] Specifically, although the candidate sequence set has been subjected to weight constraint in the generation process, there is still multi-solution and redundancy, and if it is directly used, it will lead to uneven performance of the current waveform in zero mean, spectrum distribution and energy consumption, so that it is difficult to ensure that the stimulation to the battery after superposition is controllable. By introducing the optimization process, a group of optimal solutions can be found in all candidate sequences, so that the current waveform not only meets the constraint condition, but also has mathematical orthogonality and physical effectiveness.

[0147] In the embodiment, the indicators of each candidate sequence can be used as the input quantity of the optimization target function, wherein the net electric quantity deviation is used to constrain the charge balance of the candidate sequence in a complete cycle, so as to ensure that there is no additional charging and discharging accumulation effect on the battery; the spectrum distribution is used to measure the energy concentration degree of the candidate sequence in the sensitive frequency band, so as to ensure that the electromagnetic compatibility constraint is met; and the energy consumption is used to evaluate the additional thermal energy accumulation caused by the candidate sequence in the time domain superposition process, so as to prevent the battery from overheating due to waveform superposition. Subsequently, the optimization algorithm performs global search under these constraint conditions, for example, through iterative search and convergence of the optimization target function, to filter out the first current sequence and the second current sequence that meet the complementary conditions.

[0148] It can be understood that in the obtained results, the first current sequence and the second current sequence not only satisfy zero mean value and net charge balance in a mathematical sense, but also have complementarity in frequency domain distribution, that is, the energy distribution of the two is staggered on the frequency axis, so that aliasing does not occur during subsequent demultiplexing. At the same time, the energy consumption of the two sequences is kept in a controllable range, so as to ensure that excessive thermal burden will not be caused after being superimposed on the floating current.

[0149] S3.2.3: modifying the first current sequence and the second current sequence through amplitude adjustment and phase rotation to make the first current sequence and the second current sequence orthogonal to each other within a preset time window;

[0150] Specifically, after obtaining the first current sequence and the second current sequence, further amplitude adjustment and phase rotation are needed, because the directly obtained sequence often only has orthogonality in a mathematical sense, but in the actual circuit injection process, limited by the output bandwidth of the inverter, the sampling frequency deviation and the noise interference, the orthogonality may be degraded. If the two sequences cannot maintain strict orthogonal relationship, mutual interference between signals during demultiplexing will occur, affecting the extraction accuracy of the voltage characteristic coefficient and the temperature characteristic coefficient.

[0151] In the embodiment, the amplitude of the first current sequence is normalized within a preset time window, so that the mean square energy is kept within a set threshold range, and then the phase rotation operation is applied to the second current sequence, so that the inner product result of the two within the preset time window tends to zero. The angle of phase rotation is not a fixed value, but is determined by iterative calculation. The orthogonality index is recalculated every time the rotation is performed, until the set threshold is met. In order to avoid energy mutation caused by phase adjustment, fine correction of the amplitude is also needed during the rotation process, so as to ensure the stability of the overall energy level.

[0152] S3.3: superimposing the symmetric current sequence to the floating current of the navigation substring, and obtaining the end voltage original sequence and the temperature original sequence of the navigation substring according to the sampling time sequence synchronized with the symmetric current sequence;

[0153] Specifically, the symmetric current sequence is superimposed on the floating current of the navigation substring, so as to apply a known, controllable and repeatable excitation signal to the battery module without changing the floating working state. Since the battery module is in a stable voltage maintenance interval in the floating state, the baseline current changes little at this time, so the symmetric current sequence is superimposed to introduce a slight disturbance while maintaining the power balance, and this disturbance can exactly stimulate the differential response characteristics of the battery in voltage and temperature. The symmetric current sequence is used because its positive and negative half cycles cancel each other out in terms of power, do not cause cumulative deviation of the battery capacity, and its frequency spectrum characteristics have been optimized in the foregoing steps to effectively avoid electromagnetic interference and excessive heat effects, thereby ensuring that the battery is in a safe state during actual execution.

[0154] In the embodiment, the superimposition operation is realized by a current control module of the battery management system, which superimposes the input signal according to a predetermined amplitude and time sequence on the basis of the baseline floating current, ensures consistency of the entire superimposition process with the bus voltage, and does not cause significant fluctuations in the bus. In signal acquisition, the voltage and temperature sensors are strictly synchronized with the sampling timing of the symmetric current sequence, that is, the corresponding voltage and temperature sampling is triggered at each sequence element injection time. This synchronization ensures that the acquired signal and the excitation signal have a one-to-one correspondence in time.

[0155] S3.4: Orthogonal demultiplexing and matched filter array projection are performed on the terminal voltage original sequence and the temperature original sequence to obtain a voltage characteristic coefficient vector and a temperature characteristic coefficient vector corresponding to the spectral basis;

[0156] Specifically, orthogonal demultiplexing and matched filter array projection are performed on the terminal voltage original sequence and the temperature original sequence, in order to separate the response components corresponding to different excitation sequences in the mixed response signal, and project them into the predefined spectral basis space to obtain the characteristic coefficient vector which can quantitatively characterize the battery characteristics. Since the current sequence injected in the foregoing steps has orthogonality in design, the voltage and temperature response signals can be separated without interfering with each other through the demultiplexing process, so that the response corresponding to each sequence can be recovered independently. Even if the battery responds to multiple sequences at the same time, there will be no superposition confusion in the calculation, ensuring the accuracy and stability of feature extraction.

[0157] In one example, orthogonal demultiplexing and matched filter array projection are performed on the terminal voltage original sequence and the temperature original sequence, including:

[0158] S3.4.1: The terminal voltage original sequence and the temperature original sequence are decomposed in the characteristic space composed of the spectral basis to obtain a plurality of independent response components;

[0159] Specifically, after obtaining the terminal voltage original sequence and the temperature original sequence, it is necessary to decompose them in the feature space constituted by the spectral basis, because the response of the pilot substring under the bus micro ripple disturbance is not the embodiment of a single frequency or a single physical mechanism, but is superimposed with different frequency components and dynamic processes. If decomposition is not performed, different frequency components existing in the voltage and temperature signals will be coupled with each other, which is easy to cause distortion or misjudgment in the subsequent feature extraction process. Therefore, by decomposing in the spectral basis space, the complex original signal can be separated into several independent response components, and the orthogonality and independence between different components are maintained, so that the subsequent processing process can analyze the signal features corresponding to each type of physical mechanism.

[0160] In the present embodiment, the calculation of the spectral basis has been described in the foregoing, and will not be repeated here. By projecting the original voltage sequence and the temperature sequence on the spectral basis, a set of orthogonal components can be obtained. Each component corresponds to the response of a certain frequency range or a certain dynamic process, for example, the high-frequency component reflects the direct coupling of the fast electrochemical reaction and the bus ripple, the medium-frequency component embodies the battery polarization effect, and the low-frequency component reflects the slow dynamic changes such as temperature rise and heat conduction.

[0161] Further, in the specific implementation process, in order to ensure the stability of the decomposition and the independence between the components, a decomposition method based on orthogonalization constraint can be used, for example, by normalizing and orthogonalizing the spectral basis to ensure that the components do not interfere with each other in numerical operation. In addition, historical operation data can also be combined to statistically test and remove redundant components, so as to avoid noise or abnormal disturbance components from entering the subsequent feature extraction process.

[0162] S3.4.2: Calculate the projection matrix by the bus micro ripple data recorded in the historical flight tasks, and take the current bus micro ripple data as the projection operator;

[0163] Specifically, the bus micro ripple disturbance received by the pilot substring under different task conditions is not completely consistent. If only the current ripple data is relied on for feature extraction, it is easy to be affected by accidental interference or abnormal fluctuations, resulting in distortion of the feature parameters. In order to solve this problem, a large amount of bus micro ripple data accumulated in a large number of historical flight tasks is needed as a reference sample, and a robust projection matrix is constructed based on these data, so as to provide a unified reference system for subsequent feature extraction. At the same time, the current bus micro ripple data is introduced as a projection operator, which can ensure that the calculation result contains both historical statistical rules and real-time reflection of the disturbance characteristics of the current task.

[0164] In the embodiment, the construction process of the projection matrix includes extracting typical micro ripple feature sequences from historical flight data, and screening representative frequency patterns through correlation analysis and energy distribution statistics. Subsequently, these patterns are taken as reference benchmarks to calculate the projection matrix covering the main disturbance characteristics. The projection matrix can arrange different response components in order, so that consistency with historical data can be maintained during subsequent projection calculation. Meanwhile, the current bus micro ripple data is introduced as a projection operator, which projects the original response components to the space defined by the matrix during actual calculation, thereby ensuring that the results can accurately correspond to the current disturbance state.

[0165] Further, in order to ensure the applicability of the projection matrix in long-term operation, dynamic updating logic can be used, that is, after completing a certain number of flight tasks, the matrix parameters are corrected through statistical learning of new data to avoid matrix failure due to environmental changes or battery aging. In addition, a weighting strategy can be introduced in the projection process to give higher weight to historical samples closer to the current task working condition, thereby improving the adaptability of the projection calculation to the real environment.

[0166] S3.4.3: mapping the response components corresponding to the terminal voltage original sequence and the temperature original sequence into voltage feature coefficient vectors and temperature feature coefficient vectors corresponding to the spectral basis one by one in combination with the projection matrix and the projection operator;

[0167] Specifically, if the voltage and temperature response components obtained after decomposition are not mapped, they are still numerical values distributed in the original signal space, lack a unified quantitative expression, and are not convenient for comparison between different frequency components and subsequent feature correlation analysis. Therefore, it is necessary to map these components through the joint action of the projection matrix and the projection operator to form feature coefficient vectors corresponding to the spectral basis, so that different voltage responses and temperature responses can be characterized in the same coordinate system, thereby forming consistent and comparable feature parameters.

[0168] In the embodiment, the voltage and temperature response components are first input into the projection matrix constructed by historical data, and the current bus micro ripple data is used as a projection operator to transform them. After this process, each component is mapped to the direction corresponding to the spectral basis, and a feature coefficient is output. The feature coefficient vector not only preserves the energy distribution of the original response in different frequency bands, but also eliminates the coupling between different components through matrix constraint, so that the voltage feature and the temperature feature can be expressed in the form of a vector. For example, the voltage feature coefficient vector can reflect the change amplitude of the polarization impedance at a certain frequency, and the temperature feature coefficient vector can reveal the heat accumulation characteristics at that frequency.

[0169] S3.5: processing the voltage feature coefficient vector and the temperature feature coefficient vector through compressive sensing to obtain a voltage holding transient kernel and a thermal readiness transient kernel, wherein the voltage holding transient kernel represents a voltage holding capability of the pilot substring, and the thermal readiness transient kernel represents a temperature rising accessibility of the pilot substring;

[0170] Specifically, after the voltage feature coefficient vector and the temperature feature coefficient vector are processed through the foregoing steps, the dimensions of the vectors are usually high, and there are a large number of redundant components irrelevant to the transient behavior of the battery. If these vectors are directly used for prediction, not only the calculation overhead is too large, but also the key mode of the transient behavior is submerged by noise, resulting in deviation of the evaluation result. Therefore, it is necessary to map the original high-dimensional feature vector to a low-dimensional sparse expression by using the compressive sensing technology, so as to extract the transient kernel capable of accurately representing the voltage holding capability and the temperature rising accessibility.

[0171] In the embodiment, the voltage feature coefficient vector and the temperature feature coefficient vector are processed through sparse coding respectively. First, a sparse basis dictionary related to the transient mechanism is established, which is obtained through iterative training of a large amount of historical operation data and can cover the evolution mode of the typical voltage holding process and the thermal response process. Then, the voltage feature coefficient vector is projected onto the dictionary, and the voltage holding transient kernel is obtained through sparse reconstruction. The voltage holding transient kernel represents the ability of the voltage to remain stable after the external excitation is removed in the time domain. Similarly, the temperature feature coefficient vector is processed in the same way to obtain the thermal readiness transient kernel, which reflects the accessibility of the battery module to rise in temperature after being disturbed and gradually approach a stable value.

[0172] In one example, the processing the voltage feature coefficient vector and the temperature feature coefficient vector through compressive sensing comprises:

[0173] S3.5.1: constructing a measurement matrix according to the bus micro ripple data in the historical flight mission, and projecting the voltage feature coefficient vector and the temperature feature coefficient vector into an observation domain corresponding to the measurement matrix;

[0174] S3.5.2: performing sparse representation of the projected observation vector under a preset sparse dictionary, wherein the sparse dictionary is calculated through a spectral basis;

[0175] S3.5.3: according to the sparse representation, solving an optimal sparse coefficient vector through a norm-constrained sparse recovery algorithm to reconstruct dynamic response signals corresponding to the voltage and the temperature;

[0176] S3.5.4: extracting transient features of the dynamic response signals to obtain the voltage holding transient kernel and the thermal readiness transient kernel.

[0177] S3.6: performing core playback on a standard excitation template corresponding to the symmetric current sequence by the voltage holding transient core and the thermal readiness transient core to obtain a voltage holding prediction trajectory and a temperature rise prediction trajectory, determining a maximum voltage excursion, a recovery slope and a holding time length in the voltage holding prediction trajectory as voltage responses, and determining a time to reach a micro-thermal window, a steady-state temperature difference and an overshoot in the temperature rise prediction trajectory as temperature responses;

[0178] Specifically, the voltage holding transient core and the thermal readiness transient core only provide the intrinsic ability of the battery in the transient process, but to convert it into observable and quantifiable response indicators, it is also necessary to simulate and deduce in combination with the actual current excitation. If directly dependent on the original operating conditions for evaluation, the results may lack consistency due to the variability of the operating conditions, so a standardized excitation template needs to be established, and the transient core is combined with the template by core playback to generate a standardized prediction trajectory.

[0179] In this embodiment, first, a symmetric current sequence is constructed as a standard excitation template, which is composed of current waveforms with symmetric amplitude and balanced time, which can fully stimulate the voltage and temperature responses of the battery without introducing bias. Then, the voltage holding transient core is convolved with the template to obtain a voltage holding prediction trajectory, which completely describes the excursion, recovery and holding process of the voltage under standard excitation. By analyzing the trajectory, quantifiable indicators such as maximum voltage excursion, recovery slope and holding time can be extracted, which reflect the polarization degree, recovery ability and potential stability level of the battery, respectively. Similarly, the thermal readiness transient core is played back with the standard template to obtain a temperature rise prediction trajectory, and parameters such as time to reach a micro-thermal window, steady-state temperature difference and overshoot can be obtained by analyzing the temperature rise prediction trajectory, which are used to characterize the reaction speed, temperature balance and safety margin of the battery under thermal disturbance.

[0180] In one example, the determining the dynamic internal resistance and the polarization degree of the navigation substring includes:

[0181] S3.1: dividing the maximum voltage excursion in the voltage holding prediction trajectory by the effective amplitude of the symmetric current sequence to obtain the dynamic internal resistance;

[0182] Specifically, the dynamic internal resistance reflects the voltage sensitivity of the battery to current disturbance under transient excitation, and is an important indicator for measuring the health and power response capability of the substring. If only relying on static internal resistance measurement, the real characteristics of the battery under fast dynamic conditions cannot be reflected, so it is necessary to calculate the ratio of the maximum voltage excursion in the voltage holding prediction trajectory to the current excitation amplitude to extract the internal resistance characteristics more consistent with dynamic conditions. Avoiding the interference of environmental temperature fluctuations and balancing circuit noise on static measurement, while improving the sensitivity and reliability of the performance determination of the substring.

[0183] In the embodiment, the maximum offset value of the voltage curve in the initial response stage is selected by point-by-point scanning of the voltage holding prediction trajectory obtained by compressive sensing, and the effective amplitude of the normalized symmetric current sequence is combined for ratio processing, so as to obtain the dynamic internal resistance value. Since the prediction trajectory has been projected and filtered by the spectral basis, high-frequency noise and low-frequency drift can be effectively eliminated, so that the dynamic internal resistance obtained by ratio calculation is more stable, and is suitable for battery state determination under different flight task working conditions.

[0184] S3.2: fitting the recovery slope and the holding time length in the voltage holding prediction trajectory to obtain a polarization time constant;

[0185] Specifically, after the battery experiences current excitation, the voltage will deviate, and then gradually recover in the subsequent holding stage, and the recovery speed depends on the polarization effect of the battery. By fitting the recovery slope and the holding time length, the polarization time constant can be obtained, which can represent the migration speed of the electrochemical reactant and the interface charge accumulation characteristics in the substring. If a single voltage recovery point is directly used as a determination standard, it is easy to cause distortion of the result due to transient fluctuations or measurement noise, and through the fitting processing of the slope and the time length, a more robust polarization characteristic quantity can be obtained.

[0186] In the embodiment, the voltage holding prediction trajectory is divided into an initial recovery zone and a steady-state approaching zone, the slope information of the voltage change with time is extracted respectively, and the approximate exponential decay curve is obtained by the least square fitting method, and then the corresponding polarization time constant is extracted.

[0187] S3.3: mapping the polarization time constant and the steady-state temperature difference to determine the initial polarization degree, and correcting the initial polarization degree in combination with the overshoot amount in the temperature rise prediction trajectory to obtain the polarization degree;

[0188] Specifically, the polarization time constant alone can only reflect the kinetic characteristics of the electrochemical process, but cannot comprehensively reflect the polarization intensity, so the temperature dimension needs to be introduced. The steady-state temperature difference of the battery in the holding stage can reflect the heat accumulation condition, and has a coupling relationship with the electrochemical polarization process. By mapping the polarization time constant and the steady-state temperature difference, the initial polarization degree of the substring can be obtained. If it is only stopped at the initial mapping stage, the polarization strengthening caused by the transient thermal effect may still be underestimated, so it is still necessary to combine the overshoot amount in the temperature rise prediction trajectory for secondary correction, so as to ensure that the determination of the polarization degree is more consistent with the actual operating condition.

[0189] In the embodiment, firstly, a bivariate mapping table of polarization time constant and steady-state temperature difference is established, and the mapping relationship is obtained by regression of measured data recorded in historical flight missions. Then, the time constant and steady-state temperature difference obtained by the current test are substituted into the mapping table to output the initial polarization degree. Subsequently, the maximum overshoot of the temperature rise prediction trajectory before reaching the steady state is extracted and used as a correction factor to nonlinearly adjust the initial polarization degree, thereby obtaining the final polarization degree.

[0190] In one example, the switching criteria include satisfying a zero voltage or zero current condition when the pilot sub-string is connected to the bus.

[0191] In one example, the grouping and connecting of the main sub-strings to the bus by the soft start strategy includes:

[0192] S4.1: Grouping the main sub-strings according to the voltage change rate, and performing bus voltage alignment and current limiting pre-charging before each group is connected;

[0193] In the embodiment, when grouping each main sub-string, the current end voltage of each main sub-string is first collected, and the change rate of the voltage relative to the bus voltage is calculated. The main sub-strings with similar change rates are divided into the same group. Subsequently, before the group is connected, the end voltage of the group sub-string is aligned with the bus voltage by a pre-charging circuit, and a current limiting resistor or a soft charging mode of a DC converter is used to gradually charge the capacitance of the group sub-string to a level close to the bus, so as to avoid excessive instantaneous current. At the same time, the voltage and current during the pre-charging process are monitored in real time to ensure that the group sub-string is in a safe voltage range before entering the bus.

[0194] S4.2: During the connection process, the current change rate is limited by adjusting the duty cycle of the bidirectional DC converter, and the voltage, temperature and current stability of the main sub-strings in the group are monitored;

[0195] In the embodiment, when a group is ready to be connected, the controller sets the duty cycle change curve of the bidirectional DC converter according to the real-time measured voltage difference and the preset current rising slope, so that the current gradually increases in a linear or gradual manner instead of jumping instantaneously. At the same time, the voltage balance of each main sub-string in the group and the temperature change are monitored. Once it is detected that the voltage of a certain main sub-string decreases too fast or the temperature rises too fast, the increase of the duty cycle is appropriately delayed until the sub-string recovers to stable.

[0196] S4.3: When the current group reaches the preset steady-state condition, the next group is connected, and all main sub-strings are connected to the bus;

[0197] In the present embodiment, the so-called steady-state conditions include: the fluctuation amplitude of the pack current is below a set threshold, the voltage difference among the sub-strings within the pack gradually converges to a safe range, the temperature rise rate within the pack keeps at a normal gradient, etc., which are not limited herein. When it is detected that all the conditions are met, the controller issues an instruction to access the next pack, and repeats the steps of voltage alignment, current-limiting pre-charging and duty cycle smoothing control, until all the main sub-strings are safely integrated into the bus in turn.

[0198] In one example, the present application provides an intelligent charge-discharge management system for an aviation battery, which comprises:

[0199] a battery management module for determining the pilot sub-string and the main sub-strings, and performing charge-discharge control strategies;

[0200] a signal modulation module for superimposing a symmetric current sequence in the float current, and collecting the voltage response and temperature response of the pilot sub-string;

[0201] a coordination control module for determining the dynamic internal resistance and polarization degree according to the voltage response and temperature response, and accessing the bus according to switching criteria and soft-start strategies when the main power fails.

[0202] Although the embodiments of the present application have been shown and described above, it should be understood that the above-mentioned embodiments are exemplary and should not be construed as limiting the present application, and those of ordinary skill in the art can make changes, modifications, replacements and variations to the above-mentioned embodiments within the scope of the present application.

Claims

1. A smart charging and discharging management method for aviation batteries, applied to flight control batteries, characterized in that, The flight control battery includes a busbar connected to the flight control system and multiple power supply modules, and the method includes: Through the control of the battery management system, at least one power supply module in the flight control battery is identified as the navigation substring, and the remaining power supply modules are identified as the main substring. When the flight control battery is in a float charging state, for the pilot substring, a zero-mean symmetrical current waveform is superimposed on the float charging current, and the voltage response and temperature response are calculated. The dynamic internal resistance and polarization of the pilot substring are determined based on the voltage and temperature responses, and the temperature of the pilot substring is adjusted to a preset micro-heat window and the voltage of the pilot substring is adjusted to the voltage range of the busbar through closed-loop control. When the main power supply fails, the pilot substring is connected to the bus according to the switching criteria to maintain the bus voltage, and the main substring is grouped and connected to the bus through a soft-start strategy. The soft-start strategy is used to limit the rate of change of current during the connection of the main substring. For the pilot substring, a zero-mean symmetrical current waveform is superimposed on the floating charge current, and the voltage and temperature responses are calculated, including: Acquire bus micro-ripple data recorded in historical flight missions; Principal component decomposition is performed on the bus micro ripple data, and the first K spectral bases are selected to obtain a symmetric sequence. Under the condition of satisfying the preset constraints, the symmetric sequence is processed by the optimization algorithm to obtain two mutually orthogonal symmetric current sequences with zero net charge. The constraints include electromagnetic compatibility constraints and micro-thermal energy consumption constraints. The optimization objective function of the optimization algorithm is calculated using the bus micro ripple data recorded in historical flight missions. The symmetrical current sequence is superimposed on the floating charge current of the pilot substring, and the original terminal voltage sequence and original temperature sequence of the pilot substring are obtained according to the sampling timing synchronized with the symmetrical current sequence. The original terminal voltage sequence and the original temperature sequence are orthogonally demultiplexed and projected using a matched filter array to obtain the voltage characteristic coefficient vector and temperature characteristic coefficient vector corresponding to the spectral basis. The voltage characteristic coefficient vector and temperature characteristic coefficient vector are processed by compressed sensing to obtain a voltage holding transient kernel and a thermally ready transient kernel, wherein the voltage holding transient kernel represents the potential holding capability of the pilot substring, and the thermally ready transient kernel represents the temperature rise achievable capability of the pilot substring. Based on the voltage holding transient kernel and the thermally ready transient kernel, the standard excitation template corresponding to the symmetrical current sequence is replayed to obtain the voltage holding prediction trajectory and the temperature rise prediction trajectory. The maximum voltage offset, recovery slope and holding time in the voltage holding prediction trajectory are determined as the voltage response, and the time to reach the micro-thermal window, steady-state temperature difference and overshoot in the temperature rise prediction trajectory are determined as the temperature response.

2. The intelligent charging and discharging management method for aviation batteries according to claim 1, characterized in that, At least one power supply module in the flight control battery is designated as the navigator substring, including: With each power supply module electrically isolated from the bus, a voltage following closed loop of the terminal voltage to the bus reference voltage is established by a bidirectional DC converter connected to the corresponding power supply module, wherein the voltage following closed loop uses the duty cycle of the bidirectional DC converter as the control quantity. Based on the voltage tracking closed loop, the voltage tracking error and control quantity change of each power supply module are calculated when the waveform is transformed into a unified reference waveform. The power supply modules whose tracking error is less than or equal to the first threshold and whose control quantity change is less than or equal to the second threshold within the preset time window are identified as the pilot substring.

3. The intelligent charging and discharging management method for aviation batteries according to claim 1, characterized in that, Under the condition of satisfying preset constraints, the symmetrical sequence is processed by an optimization algorithm to obtain two mutually orthogonal symmetrical current sequences with zero net charge, including: The symmetric sequence is decomposed according to frequency components and time-domain components, and the decomposed vectors are weighted by preset weights to obtain a candidate sequence set, wherein the weights are calculated based on electromagnetic compatibility constraints and micro-thermal energy consumption constraints. The candidate sequence set is input into a preset optimization algorithm to calculate the net charge deviation, spectral distribution and energy consumption of the candidate sequences, so as to obtain the first current sequence and the second current sequence. The first and second current sequences are corrected by amplitude adjustment and phase rotation so that the first and second current sequences are orthogonal to each other within a preset time window.

4. The intelligent charging and discharging management method for aviation batteries according to claim 1, characterized in that, The process of orthogonally demultiplexing and projecting a matched filter array onto the original terminal voltage and temperature sequences includes: The original terminal voltage sequence and the original temperature sequence are decomposed in the feature space formed by the spectral basis to obtain multiple independent response components. The projection matrix is ​​calculated using bus micro-ripple data recorded in historical flight missions, and the current bus micro-ripple data is used as the projection operator. By combining the projection matrix and projection operator, the response components corresponding to the original voltage and temperature sequences are mapped to voltage and temperature characteristic coefficient vectors that correspond one-to-one with the spectral basis.

5. The intelligent charging and discharging management method for aviation batteries according to claim 1, characterized in that, The process of processing the voltage characteristic coefficient vector and temperature characteristic coefficient vector through compressed sensing includes: A measurement matrix is ​​constructed based on bus micro-ripple data from historical flight missions, and the voltage characteristic coefficient vector and temperature characteristic coefficient vector are projected onto the observation domain corresponding to the measurement matrix. The projected observation vector is sparsed under a predefined sparse dictionary, wherein the sparse dictionary is calculated using spectral basis. Based on the sparsified representation, the optimal sparse coefficient vector is solved by a norm-constrained sparse recovery algorithm to reconstruct the dynamic response signals corresponding to voltage and temperature. The transient features of the dynamic response signal are extracted to obtain the voltage holding transient kernel and the thermally ready transient kernel.

6. The intelligent charging and discharging management method for aviation batteries according to claim 1, characterized in that, The determination of the dynamic internal resistance and polarization of the pilot substring includes: The dynamic internal resistance is obtained by dividing the maximum voltage offset in the voltage holding prediction trajectory by the effective amplitude of the symmetrical current sequence. The polarization time constant is obtained by fitting the recovery slope and holding time in the voltage holding predicted trajectory. The initial polarization degree is determined by mapping the polarization time constant to the steady-state temperature difference, and then the initial polarization degree is corrected by combining the overshoot in the temperature rise prediction trajectory to obtain the polarization degree.

7. The intelligent charging and discharging management method for aviation batteries according to claim 1, characterized in that, The switching criteria include meeting the zero voltage or zero current condition when the pilot substring is connected to the bus.

8. The intelligent charging and discharging management method for aviation batteries according to claim 1, characterized in that, The step of grouping the main substrings into the bus using a soft-start strategy includes: The main substrings are grouped according to the voltage change rate, and bus voltage alignment and current-limiting pre-charging are performed before each group is connected. During the integration process, the current change rate is limited by adjusting the duty cycle of the bidirectional DC-DC converter, and the voltage, temperature and current stability of the main substring within the group are monitored. Once the current group reaches the preset steady-state condition, the next group is connected until all main substrings are merged into the bus.

9. An intelligent charging and discharging management system for aviation batteries, used to implement the intelligent charging and discharging management method for aviation batteries as described in any one of claims 1-8, characterized in that, The system includes: The battery management module is used to determine the pilot substring and the main substring, and to execute the charge and discharge control strategy; The signal modulation module is used to superimpose a symmetrical current sequence on the floating charge current and to acquire the voltage and temperature responses of the pilot substring. The coordination control module is used to determine the dynamic internal resistance and polarization degree based on the voltage response and temperature response, and to connect to the bus according to the switching criteria and soft-start strategy when the main power supply fails.

Images