An aircraft battery state adaptive detection method, system and aircraft

Abstract

The application provides an aircraft battery state adaptive detection method, system and aircraft. The method comprises the following steps: obtaining task planning information of the aircraft, performing virtual deduction on the flight task by using a pre-constructed battery-aircraft digital twin model, and generating a theoretical state evolution curve of each stage of the flight; in the flight process, real-time collection of flight state data and battery parameter data, multi-time scale joint estimation of the state of charge, power state and energy state of the battery, and obtaining of real-time state estimation values; comparison of the theoretical state evolution curve and the real-time state estimation values, prospective task feasibility determination, and determination including energy feasibility determination and power feasibility determination; if the result of the prospective task feasibility determination does not satisfy a preset condition, a warning signal and / or a flight strategy adjustment instruction are generated. Through the above method, continuous optimization and adaptive improvement of detection accuracy are realized.

Description

Technical Field

[0001] This application relates to the field of aircraft battery status detection technology, and in particular to an adaptive detection method for aircraft battery status, an adaptive detection system for aircraft battery status, and an aircraft. Background Technology

[0002] With the widespread application of electric vertical takeoff and landing aircraft and drones in logistics delivery, urban air travel, power line inspection and other fields, battery status detection and safety assessment of aircraft have become core issues to ensure flight safety.

[0003] Existing methods for detecting the state of aircraft batteries have the following main shortcomings: (i) The utilization of working conditions is superficial and lacks forward-looking understanding. While some existing technologies mention flight condition classification, they merely use these conditions as model input labels or simply switch detection thresholds, failing to establish an explicit mapping relationship between "flight conditions" and "internal physical stress of the battery." More importantly, existing detection methods are all "response-based"—detecting only what condition is currently in use, lacking forward-looking cognitive capabilities based on flight mission planning.

[0004] (ii) The detection dimension is singular and does not consider multi-state coupling. Existing technologies primarily focus on estimating the State of Charge (SOC), with less emphasis on the joint estimation of the State of Power (SOP) and State of Energy (SOE). For aircraft, even with a sufficient SOC, if the SOP is insufficient to support upcoming high-power conditions such as climbs, flight safety risks still exist. Single-dimensional testing cannot comprehensively assess the battery's capability margin in complex flight missions.

[0005] (iii) The model is static and fixed, lacking a closed-loop update mechanism. In existing technologies, battery testing models remain fixed once deployed, unable to be updated online or adaptively adjusted based on actual data accumulated during flight. Battery aging characteristics vary depending on operating conditions; the aging trajectory of the same battery model may differ significantly across different aircraft, making it difficult for static models to adapt to these individual differences. Summary of the Invention

[0006] To address the aforementioned technical problems, this application proposes an adaptive detection method for aircraft battery status, an adaptive detection system for aircraft battery status, and an aircraft.

[0007] To address the aforementioned technical problems, this application proposes an adaptive detection method for aircraft battery status, comprising the following steps: The mission planning information of the aircraft is obtained, and the flight mission is virtually simulated using a pre-built battery-aircraft digital twin model to generate theoretical state evolution curves for each stage of flight. The theoretical state evolution curves include at least the state of charge evolution curve, the power state evolution curve, and the energy state evolution curve. During flight, flight status data and battery parameter data are collected in real time, and the battery's state of charge, state of power, and state of energy are jointly estimated at multiple time scales to obtain real-time state estimates. The theoretical state evolution curve is compared with the real-time state estimate to make a forward-looking task feasibility determination, which includes energy feasibility determination and power feasibility determination. If the result of the feasibility assessment of the prospective mission does not meet the preset conditions, an early warning signal and / or flight strategy adjustment instruction are generated, and the real-time data is fed back to the digital twin model to update the theoretical state evolution curve. The flight strategy adjustment instructions include one or more of the following: reducing cruise speed, reducing climb rate, modifying waypoint order, or returning to base.

[0008] The battery-aircraft digital twin model includes an electrical model, a thermal model, and a capacity model. The electrical model is used to simulate the voltage response and power output characteristics of a battery under different operating conditions. The thermal model is used to simulate the temperature rise characteristics of the battery during flight; The capacity model is used to simulate the capacity decay characteristics of a battery.

[0009] The virtual simulation includes: Based on the waypoint sequence, altitude profile, and velocity profile in the mission planning information, predict the operating condition sequence for each stage of flight; The theoretical state evolution curve is generated by simulating the evolution of battery state parameters based on the operating condition sequence.

[0010] The joint estimation across multiple time scales includes: The state of charge and the state of energy are estimated at a first time scale, which is 5-10 seconds. The power state is estimated on a second time scale, which is 100-500 milliseconds; Wherein, the first time scale is larger than the second time scale.

[0011] The joint estimation of the battery's state of charge, state of power, and state of energy across multiple time scales is achieved using either the extended Kalman filter algorithm or the unscented Kalman filter algorithm.

[0012] The energy feasibility determination includes: Compare the state of charge or energy state in the real-time state estimate with the predicted value at the corresponding moment in the theoretical state evolution curve; Determine whether the remaining energy is sufficient to complete the subsequent flight mission.

[0013] The power feasibility determination includes: Predict the peak power requirements for future operating conditions based on the aforementioned aircraft mission planning information; Compare the peak power demand with the power state in the real-time state estimate; Determine whether the battery's peak power output capability meets future operating conditions. The peak power demand is determined based on a predicted sequence of future operating conditions, which includes at least one or more of climbing, cruising, and hovering conditions.

[0014] Specifically, feeding the real-time data back to the digital twin model to update the theoretical state evolution curve includes: During flight or when the judgment result does not meet the preset conditions, the actual collected voltage, current, temperature, and SOC data will be fed back to the digital twin model. By correcting the model parameters, subsequent calculations can be made more accurate. The model parameters include at least the internal resistance value and the capacity decay coefficient.

[0015] To address the aforementioned technical problems, this application also proposes an adaptive detection system for aircraft battery status, comprising: The digital twin module is used to acquire mission planning information of the aircraft, perform virtual simulation of the flight mission using a pre-built battery-aircraft digital twin model, and generate theoretical state evolution curves for each stage of flight. The theoretical state evolution curves include at least the state of charge evolution curve, the power state evolution curve, and the energy state evolution curve. The state estimation module is used to collect flight state data and battery parameter data in real time during flight, and to perform joint estimation of the battery's state of charge, state of power and state of energy at multiple time scales to obtain real-time state estimates. The feasibility assessment module is used to compare the theoretical state evolution curve with the real-time state estimate to make a forward-looking task feasibility assessment, which includes energy feasibility assessment and power feasibility assessment. An adaptive adjustment module is used to generate an early warning signal and / or flight strategy adjustment command when the result of the feasibility determination of the prospective mission does not meet the preset conditions, and to feed the real-time data back to the digital twin module to update the theoretical state evolution curve.

[0016] To address the aforementioned technical problems, this application also proposes an aircraft, including the aforementioned adaptive detection system for aircraft battery status.

[0017] Compared with existing technologies, the beneficial effects of this application are as follows: By pre-constructing a battery-aircraft digital twin model to virtually simulate flight missions and generate theoretical state evolution curves, the detection system possesses a forward-looking understanding of battery state evolution trends before flight begins. This understanding serves as a benchmark for online detection, thereby upgrading from passive response-based detection to proactive predictive detection. By real-time acquisition of flight status data and battery parameter data, the system performs joint estimation of the battery's state of charge, power state, and energy state across multiple time scales. This allows the system to adopt appropriate estimation cycles for the changing characteristics of different states, ensuring the stability of energy state assessment and improving the response speed to transient changes in power state, thus achieving a comprehensive coupled assessment of battery energy margin and power margin. By comparing the theoretical state evolution curve with real-time state estimates to make a forward-looking mission feasibility determination, the system can identify risks in advance when energy is insufficient or power is inadequate to support subsequent flight scenarios, thereby achieving early warning of flight safety risks. By feeding real-time data back to the digital twin model to update the theoretical state evolution curve, the detection system can adjust model parameters online based on actual data accumulated during flight, adapting to the individual differences in battery aging characteristics, thereby achieving continuous optimization of detection accuracy and improvement of adaptability. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. Wherein: Figure 1 This is a flowchart illustrating the first embodiment of the adaptive detection method for aircraft battery status provided in this application; Figure 2 This is a flowchart illustrating the second embodiment of the adaptive detection method for aircraft battery status provided in this application; Figure 3 This is a flowchart illustrating the third embodiment of the adaptive detection method for aircraft battery status provided in this application; Figure 4 This is a flowchart illustrating the fourth embodiment of the adaptive detection method for aircraft battery status provided in this application; Figure 5 This is a flowchart illustrating the fifth embodiment of the adaptive detection method for aircraft battery status provided in this application; Figure 6 This is a flowchart illustrating the sixth embodiment of the adaptive detection method for aircraft battery status provided in this application; Figure 7 This is a schematic diagram of an embodiment of the adaptive detection system for aircraft battery status provided in this application; Figure 8 This is a structural schematic diagram of an embodiment of the aircraft provided in this application. Detailed Implementation

[0019] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of the embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0020] The terms “first,” “second,” “third,” “fourth,” etc. (if present) in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a particular order or sequence. It should be understood that such data can be interchanged where appropriate so that embodiments of the application described herein can be implemented, for example, in orders other than those illustrated or described herein. Furthermore, the terms “comprising” and “having,” and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0021] This application provides an adaptive detection method for aircraft battery status. Its protection boundary is to generate a theoretical benchmark by performing virtual simulation before takeoff using a digital twin model, combined with real-time estimation at multiple time scales during flight, to make a forward-looking feasibility judgment, and to establish a closed-loop feedback mechanism to update the model.

[0022] Please refer to the details. Figure 1 , Figure 1 This is a flowchart illustrating the first embodiment of the adaptive detection method for aircraft battery status provided in this application.

[0023] like Figure 1 As shown, the specific steps are as follows: Step S11: Obtain the mission planning information of the aircraft, use the pre-built battery-aircraft digital twin model to virtually simulate the flight mission, and generate theoretical state evolution curves for each stage of flight. The theoretical state evolution curves include at least the state of charge evolution curve, the power state evolution curve, and the energy state evolution curve.

[0024] In this application embodiment, the digital twin model refers to a virtual battery-aircraft coupling model constructed through an equivalent circuit model, a thermal model, and a capacity model, used to simulate the physical response of the battery under different operating conditions.

[0025] This application acquires mission planning information and utilizes a digital twin model for virtual simulation, with the aim of establishing a baseline before flight commencement. By inputting waypoints, altitude, and velocity profiles, the model simulates the operational condition sequence for each stage, thereby generating theoretical evolution curves for SOC, SOP, and SOE. This addresses the lack of forward-looking knowledge in existing technologies, upgrading detection from reactive to predictive. In essence, before takeoff, the system has already 'pre-rehearsed' the flight process, knowing how much battery power should remain and how much power it can provide at what point in time.

[0026] Step S12: During flight, flight status data and battery parameter data are collected in real time, and the battery's state of charge, state of power and state of energy are jointly estimated at multiple time scales to obtain real-time state estimates.

[0027] In this embodiment, the state of charge (SOC) refers to the percentage of remaining battery capacity, the state of power (SOP) refers to the maximum power output capability of the battery under the current environment, and the state of energy (SGE) refers to the total remaining usable energy of the battery. Multi-timescale joint estimation refers to an estimation strategy that employs different sampling periods for different state parameters.

[0028] Multi-timescale joint estimation is performed during flight. This design is based on the fact that SOC and SOE change relatively slowly, making them suitable for longer-period (e.g., second-level) estimations to smooth noise; while SOP involves transient power output and is greatly affected by rapid changes in operating conditions, requiring short-period (e.g., millisecond-level) estimations to respond promptly to high power demands. This design allows the system to both grasp overall energy margins and cope with sudden power peak demands. For example, an extended Kalman filter algorithm can be used to process SOC / SOE, while a lookup table method or online identification algorithm can be used to process SOP.

[0029] Step S13: Compare the theoretical state evolution curve with the real-time state estimate to make a forward-looking task feasibility determination, which includes energy feasibility determination and power feasibility determination.

[0030] In this embodiment, a forward-looking task feasibility assessment is performed. Its working principle involves comparing real-time estimates with theoretical curves. Energy feasibility assessment focuses on whether remaining energy can support subsequent tasks, while power feasibility assessment focuses on whether current power capacity meets future peak operating conditions. This comparison achieves a shift from a single-dimensional to a multi-dimensional coupled evaluation, ensuring not only 'having electricity' but also 'having power'.

[0031] Step S14: If the result of the feasibility determination of the prospective mission does not meet the preset conditions, generate an early warning signal and / or flight strategy adjustment command, and feed the real-time data back to the digital twin model to update the theoretical state evolution curve.

[0032] In this embodiment, a closed-loop feedback update is used. When the condition is not met, an early warning or adjustment command is generated, and real-time data is fed back to the model. Its function is to use deviations in actual flight (such as increased internal resistance leading to faster temperature rise) to correct model parameters, making subsequent inferences more accurate and adapting to the aging characteristics of the battery.

[0033] Take an eVTOL aircraft performing logistics delivery missions as an example. Before takeoff, the system acquires the mission plan: vertical takeoff, climb to 200 meters, cruise for 15 kilometers, hover and release, and return. The digital twin model generates a theoretical curve based on this, predicting that the State of Charge (SOC) during the cruise phase should be 65%. During flight, the Battery Management System (BMS) estimates SOC / State of Energy (SOE) every 5 seconds and State of Operation (SOP) every 200 milliseconds. If the current SOC is detected to be 58%, lower than the predicted value, an insufficient energy feasibility warning is triggered. Simultaneously, if it is predicted that 45 kW of power is needed to enter the climb phase in 40 seconds, while the current estimated SOP is 40 kW, a insufficient power feasibility warning is triggered. The system generates a warning at the attention level and suggests reducing the cruise speed, while simultaneously feeding the measured data back to the model to correct internal resistance parameters and update subsequent projection curves.

[0034] Furthermore, this application can use unscented Kalman filtering instead of extended Kalman filtering for state estimation; the equivalent circuit model in the digital twin model can be a second-order RC model or a higher-order model; the thermal model can be a lumped parameter thermal model or a finite element thermal model; and the capacity model can be nonlinearly corrected based on the cumulative discharge depth and operating condition type.

[0035] This application utilizes a pre-constructed battery-aircraft digital twin model to virtually simulate flight missions and generate theoretical state evolution curves. This allows the detection system to possess a forward-looking understanding of battery state evolution trends before flight begins, serving as a benchmark for online detection and thus upgrading from passive response-based detection to proactive predictive detection. By acquiring real-time flight status data and battery parameter data, the system performs joint estimation of the battery's state of charge, state of power, and state of energy across multiple time scales. This enables the system to employ appropriate estimation cycles for the changing characteristics of different states, ensuring the stability of the state of energy assessment while improving the response speed to transient changes in the state of power, thereby achieving a comprehensive coupled assessment of battery energy margin and power margin. By comparing the theoretical state evolution curves with real-time state estimates to conduct forward-looking mission feasibility assessments, the system can identify risks in advance when energy is insufficient or power is inadequate to support subsequent flight scenarios, thus achieving early warning of flight safety risks. By feeding real-time data back to the digital twin model to update the theoretical state evolution curve, the detection system can adjust model parameters online based on actual data accumulated during flight, adapting to the individual differences in battery aging characteristics, thereby achieving continuous optimization of detection accuracy and improvement of adaptability.

[0036] based on Figure 1 The adaptive detection method for aircraft battery status further defines the specific composition of the battery-aircraft digital twin model in this application. Its protection boundary clarifies that the digital twin model is composed of three coupled sub-models: an electrical model, a thermal model, and a capacity model, and defines the function of each sub-model.

[0037] Among them, the electrical model refers to a mathematical model built on the principle of equivalent circuits, used to describe the response of the internal electrochemical process of the battery to the external circuit, mainly characterizing the relationship between voltage, current and power; the thermal model refers to a mathematical model built on the principle of thermodynamics, used to describe the heat generated by the battery during charging and discharging and its transfer and dissipation process within the battery pack, characterizing the change of battery temperature with time and operating conditions; the capacity model refers to a mathematical model built on the battery aging mechanism, used to describe the characteristic that the battery's usable capacity gradually decreases with the accumulation of cycle number, depth of discharge and type of operating conditions.

[0038] Specifically, the electrical model is designed to simulate the battery's voltage response and power output characteristics under different operating conditions. This is because the load current experienced by the battery varies significantly during different flight phases of the aircraft (such as climb, cruise, and hovering), and a single static model cannot accurately reflect the voltage drop and power fluctuations during dynamic processes. The electrical model calculates the battery's internal resistance voltage drop and polarization voltage by inputting the current curves corresponding to the operating conditions, thereby outputting a precise voltage response and available power. For example, a second-order RC equivalent circuit model can be used, which includes an ohmic internal resistance and two parallel RC branches to simulate the battery's transient response and medium- to long-term polarization effects, respectively. A pseudo-two-dimensional (P2D) model based on physical mechanisms can be used as an alternative implementation to provide a more accurate simulation of the internal electrochemical state.

[0039] The thermal model is designed to simulate the temperature rise characteristics of the battery during flight. This is because the battery generates significant Joule heat during high-rate discharge (such as during eVTOL takeoff and climb phases). Increased temperature not only affects battery performance but may also trigger thermal runaway. The thermal model calculates the spatiotemporal distribution of battery temperature by inputting the heat generation power from the electrical model and considering the battery's thermal capacity, thermal resistance, and heat dissipation conditions. For example, a lumped-parameter thermal model can be used, treating the battery as a isothermal body and calculating the temperature rise using the thermal balance equation dT / dt=(Q_gen-Q_loss) / C; alternatively, a finite element thermal model can be used, considering the heat conduction paths within the battery pack. This design allows the system to predict the battery's temperature state under subsequent high-power conditions, avoiding power derating due to overheating.

[0040] The capacity model is designed to simulate the capacity decay characteristics of batteries. This is because batteries undergo irreversible aging during use, leading to a reduction in usable capacity, and the degree of aging is influenced by historical operating conditions. The capacity model corrects for the theoretical initial capacity of the battery by accumulating the depth of discharge during flight, the number of cycles, and the average temperature. For example, an empirical model based on the Arrhenius equation can be used, with temperature as an acceleration factor, combined with the number of cycles to establish a capacity decay curve. This design allows the digital twin model to reflect the individual aging trajectory of a single battery cell or battery pack, avoiding prediction bias caused by discrepancies between model aging parameters and actual conditions.

[0041] Take an eVTOL aircraft performing logistics delivery missions as an example. Before takeoff, the digital twin model loads historical test data for that batch of batteries. The electrical model uses a second-order RC model with parameters including an internal resistance of 0.5mΩ and a polarization resistance of 1mΩ. The thermal model uses a lumped parameter model with a heat capacity of 500J / K and a thermal resistance of 2K / W. The capacity model is based on the initial capacity of 22000mAh of this battery model at the time of manufacture, and is corrected by combining the cumulative discharge depth from previous flights. During simulated flight missions, the electrical model calculates the voltage drop based on the high current demand during the climb phase, the thermal model calculates the battery surface temperature based on the heat output from the electrical model, and the capacity model updates the current available capacity limit based on the cumulative discharge. These three models are coupled together: the electrical model outputs heat power to the thermal model, the thermal model outputs temperature feedback to the electrical model to correct the internal resistance, and the capacity model corrects the available energy range of the electrical model as a whole.

[0042] Furthermore, this application can use a first-order RC model instead of a second-order RC model to reduce computational complexity; the thermal model can be refined using finite element analysis software for thermal simulation; and the capacity model can be a capacity decay prediction model based on machine learning algorithms (such as LSTM neural networks), with historical charge and discharge data input to capture nonlinear aging characteristics.

[0043] This application establishes an electrical model to simulate the battery's voltage response and power output characteristics under different operating conditions. This allows the digital twin model to accurately reflect the battery's electrical performance under the complex dynamic loads of the aircraft, providing accurate voltage and power benchmarks for power state assessment and thus improving the accuracy of power feasibility determination. By establishing a thermal model to simulate the battery's temperature rise characteristics during flight, the system can predict the battery's temperature change trend under high-power conditions, avoiding performance degradation or safety hazards caused by high temperatures, thereby ensuring the battery's operational safety under extreme conditions. Furthermore, by establishing a capacity model to simulate the battery's capacity decay characteristics, the digital twin model can dynamically reflect the aging differences that occur during battery use, correcting the energy benchmark in the theoretical state evolution curve. This improves the accuracy of remaining energy state prediction and avoids energy estimation errors caused by battery aging.

[0044] based on Figure 1 The aforementioned adaptive detection method for aircraft battery status provides a specific implementation method for virtual simulation, namely, deriving the battery's workload by analyzing the kinematic parameters in the mission planning, thereby establishing an explicit mapping relationship between flight conditions and the battery's internal state.

[0045] Please refer to the details. Figure 2 , Figure 2 This is a flowchart illustrating the second embodiment of the adaptive detection method for aircraft battery status provided in this application.

[0046] like Figure 2 As shown, the specific steps are as follows: Step S21: Based on the waypoint sequence, altitude profile, and velocity profile in the mission planning information, predict the operating condition sequence for each stage of flight.

[0047] In the embodiments of this application, the operating condition sequence refers to the set of load characteristics corresponding to different stages of the flight mission (such as vertical take-off and landing, climb, cruise, hovering, descent, etc.), specifically reflected in the current magnitude, discharge rate and duration.

[0048] This application obtains waypoint sequences, altitude profiles, and velocity profiles, which are fundamental to constructing operational condition sequences. Waypoint sequences define the spatial path of flight, while altitude and velocity profiles define the aircraft's motion state in space. Based on aerodynamic principles and aircraft dynamics models, the aircraft's attitude angles, acceleration, and rotor / motor speeds are directly related to its current speed and altitude. By inputting these kinematic parameters into the aircraft's power consumption calculation model, the required output power of the motors at each time step can be accurately calculated, thereby deducing the current and discharge rate required by the battery. This derivation process from "flight mission" to "battery load" solves the problems of shallow operational condition utilization and lack of forward-looking understanding in existing technologies, enabling the system to predict the specific physical stress the battery will experience before flight.

[0049] Step S22: Based on the operating condition sequence, simulate the evolution process of battery state parameters and generate the theoretical state evolution curve.

[0050] In the embodiments of this application, the theoretical state evolution curve refers to the ideal reference trajectory of the battery's state of charge, power state, and energy state changing over time, which is simulated in a virtual environment based on a digital twin model.

[0051] This application simulates the evolution of battery state parameters based on a sequence of operating conditions, utilizing a coupled simulation of the electrical, thermal, and capacity models within a digital twin model. The electrical model calculates the voltage response and SOC change based on the input current curve; the thermal model calculates the battery temperature rise based on Joule heating (proportional to the square of the current); and the capacity model corrects for remaining capacity based on the accumulated depth of discharge and temperature. These three models work together to simulate the time-varying trajectories of SOC, SOP, and SOE under a specific sequence of operating conditions, i.e., the theoretical state evolution curves. This simulation process transforms the abstract flight mission into concrete changes in the internal physical quantities of the battery, providing a precise benchmark for subsequent real-time monitoring.

[0052] Take an eVTOL aircraft performing logistics delivery missions as an example. Before takeoff, the system obtains the mission plan: vertically ascend to an altitude of 200 meters from the takeoff point, then cruise for 15 kilometers at a speed of 25 meters per second, hover for 30 seconds to drop cargo upon reaching the target point, and finally return and land at a speed of 20 meters per second. Based on the waypoint sequence and altitude and speed profiles, the system predicts the operating condition sequence for each stage of flight: the takeoff stage is a high-current short-term discharge, the climb stage is a medium-to-high-current continuous discharge, the cruise stage is a medium-current continuous discharge, the hovering stage is a medium-current short-term discharge, and the return stage is similar to the cruise stage. Subsequently, the system inputs this operating condition sequence into a digital twin model. For example, during the climb stage, the model, based on a preset current curve and the current ambient temperature, simulates the temperature rise of the battery due to internal resistance heating during the climb, as well as the resulting voltage drop and SOC consumption rate, thereby generating SOC evolution curves and SOP evolution curves with temperature compensation. A second-order RC equivalent circuit model can be used as the electrical model to capture the dynamic voltage response of the battery; a lumped parameter thermal model can be used as the thermal model to quickly calculate the average temperature rise of the battery pack; and an empirical model based on the Arrhenius equation can be used as the capacity model to correct the effect of temperature on capacity.

[0053] Furthermore, this application may employ unscented Kalman filtering or extended Kalman filtering algorithms to assist in the smoothing of the operating condition sequence; the prediction of the operating condition sequence may be based on searching pre-stored wind tunnel test data or real-time dynamic calculations of the flight control system; the electrical model in the digital twin model may adopt an equivalent circuit model or an electrochemical model; the thermal model may adopt a lumped parameter thermal model or a finite element thermal model; and the capacity model may be nonlinearly corrected based on the cumulative discharge depth and operating condition type.

[0054] This application predicts the operating condition sequence for each stage of flight based on waypoint sequences, altitude profiles, and velocity profiles from mission planning information. This allows the system to transform abstract flight missions into specific battery load characteristics, establishing an explicit mapping relationship between flight operating conditions and internal battery physical stress. This upgrades the system from passive response detection to proactive, mission-plan-based understanding. By simulating the evolution of battery state parameters based on the operating condition sequence and generating theoretical state evolution curves, the system can predict the trends of battery state of charge, power state, and energy state changes in subsequent stages before flight begins. This provides a high-precision benchmark for real-time monitoring, enabling proactive prediction of battery state evolution trends.

[0055] based on Figure 1The adaptive detection method for aircraft battery status described herein uses differentiated time scales to estimate different state parameters in multi-state joint estimation. Specifically, the estimation period for the state of charge and energy is limited to 5-10 seconds, and the estimation period for the power state is 100-500 milliseconds, with the former being longer than the latter.

[0056] Please refer to the details. Figure 3 , Figure 3 This is a flowchart illustrating the third embodiment of the adaptive detection method for aircraft battery status provided in this application.

[0057] like Figure 3 As shown, the specific steps are as follows: Step S31: Estimate the state of charge and the state of energy on a first time scale of 5-10 seconds.

[0058] In the embodiments of this application, the state of charge refers to the proportion of the battery's current remaining charge to its total capacity; the state of energy refers to the total electrical energy that the battery can currently release.

[0059] The state of charge (POC) and state of energy (SEE) are estimated using a first timescale of 5-10 seconds. This is because POC and SEE are thermodynamic cumulative parameters of the battery, affected by current integrals, and change relatively slowly, being highly sensitive to transient noise. If the sampling frequency is too high, sensor noise can easily be introduced, leading to estimation drift; if the frequency is too low, it cannot reflect energy consumption trends in a timely manner. A 5-10 second period strikes a balance between filtering high-frequency noise and maintaining data timeliness. An extended Kalman filter algorithm can be used on this timescale, employing the average values ​​of voltage, current, and temperature as input to smooth the data and obtain a stable energy margin assessment.

[0060] Step S32: Estimate the power state on a second time scale, which is 100-500 milliseconds; wherein the first time scale is longer than the second time scale.

[0061] In the embodiments of this application, the power state refers to the maximum electrical power that the battery can continuously or at its peak output under specific operating conditions; the first time scale refers to the sampling or update period used to estimate the state of charge and the state of energy; the second time scale refers to the sampling or update period used to estimate the power state.

[0062] The power state is estimated using a second timescale of 100-500 milliseconds. This is because power demands fluctuate drastically within a very short timeframe during flight operations such as climb and hover transitions. The power state directly depends on the battery's internal resistance, temperature, and current state of charge (SOC), parameters that respond transiently to load changes. The short 100-500 millisecond timescale captures voltage drops and power output limits under millisecond-level load transitions, ensuring the system accurately determines whether the battery has sufficient power margin before entering high-power operating conditions. A lookup table method based on an equivalent circuit model or an online parameter identification algorithm can be used to recalculate the battery's maximum permissible charge / discharge current at the current SOC and temperature every 100-500 milliseconds, and then convert this into the power state.

[0063] Regarding the logical relationship between the first time scale and the second time scale, its working principle is a hierarchical estimation strategy. Slow variables (SOC / SOE) provide a macroscopic energy benchmark, while fast variables (SOP) provide microscopic power response capabilities. This design makes system resource allocation more rational, eliminating the need for complex full updates of SOC / SOE in every estimation. Instead, based on short-cycle SOP estimations, SOC / SOE is synchronously corrected every 5-10 seconds, reducing computational load while ensuring the real-time performance of key safety indicators (power).

[0064] Taking an eVTOL aircraft performing a logistics mission as an example. During the cruise phase, the BMS updates the SOC and SOE estimates every 5 seconds. At this time, the motor load is stable, voltage and current fluctuations are small, and the 5-second integration window is sufficient to eliminate noise and obtain an accurate prediction of the remaining range. When the aircraft is preparing to transition from cruise to climb mode, the flight controller issues a power increase command. At this time, the BMS calculates the SOP at a high frequency with a period of 200 milliseconds. Within 200 milliseconds, the system reads the latest voltage, current, and temperature, and calculates whether the current battery can provide the peak power required for climb, based on the battery internal resistance model. If the SOP estimate within 200 milliseconds is lower than the climb requirement threshold, the system immediately triggers an early warning without waiting for the SOC update after 5 seconds. In another embodiment, 100 milliseconds can be used as the SOP estimation period, which is suitable for competitive UAVs with extremely high response speed requirements; or 500 milliseconds can be used as the SOP period, which is suitable for long-endurance inspection UAVs with limited computing resources.

[0065] Furthermore, this application can use unscented Kalman filtering instead of extended Kalman filtering for state estimation on the first time scale; on the second time scale, a power prediction model based on support vector machine can be used instead of the lookup table method; the time scale can be adjusted to 5 seconds and 200 milliseconds, or 10 seconds and 500 milliseconds, as long as the first scale is greater than the second scale.

[0066] This application estimates the state of charge (SOC) and state of energy (SOE) by setting a first timescale of 5-10 seconds. This allows the system to filter out high-frequency noise from the sensors while maintaining appropriate tracking of the battery's remaining energy change trend, thus achieving stability and accuracy in SOE estimation. By setting a second timescale of 100-500 milliseconds for SOE estimation, the system can closely follow rapid changes in aircraft operating conditions and capture the battery's power output limit under transient loads in real time, enabling timely response and risk prediction for high-power demand scenarios. By making the first timescale longer than the second timescale, the system can employ a hierarchical estimation strategy, reducing the overall computational load while maintaining the stability of SOE and the real-time performance of SOE, thereby improving the efficiency and safety of multi-state joint estimation.

[0067] based on Figure 1 The adaptive detection method for aircraft battery status, as described in this application, provides a specific algorithm selection for the aforementioned multi-timescale joint estimation, namely, the extended Kalman filter algorithm or the unscented Kalman filter algorithm.

[0068] The Extended Kalman Filter (EKF) is a recursive filtering algorithm that estimates the optimal state of a dynamic system by linearizing the nonlinear system model (usually using a first-order Taylor series expansion). The Unscented Kalman Filter (UKF) is a state estimation algorithm based on the Unscented Transform. It approximates the probability distribution of the state by selecting a set of sigma points with specific weights, thereby avoiding the linearization approximation of the nonlinear model.

[0069] In aircraft battery management, the electrochemical processes within the battery exhibit highly nonlinear characteristics, such as the nonlinear relationship between internal resistance and temperature / state of charge (SOC), and the dynamic response of polarization voltage. Traditional linear filtering algorithms struggle to handle this strong nonlinearity, leading to accumulated estimation errors. Extended Kalman filtering (EPK) locally linearizes the nonlinear model using the Jacobian matrix, providing relatively accurate state estimates within an acceptable computational complexity range, making it suitable for most conventional flight conditions. Unscented Kalman filtering, on the other hand, directly handles nonlinear distributions through unscented transformations, without relying on derivative calculations. Therefore, it provides more robust estimation accuracy even with extremely high model nonlinearity or significant noise, making it particularly suitable for extreme conditions or situations where severe battery aging leads to model mismatch.

[0070] In multi-timescale estimation, whether it's second-level estimation of SOC / SOE or millisecond-level estimation of SOP, real-time processing of voltage, current, and temperature data from the BMS is required. Extended Kalman filtering uses a two-step prediction-update process to correct the previous state estimate using current measurements, thus tracking changes in the battery's internal state in real time. Unscented Kalman filtering propagates the sigma point through a nonlinear system to obtain a more accurate approximation of the state mean and covariance, thereby updating the state estimate. The introduction of these two algorithms makes joint estimation across multiple timescales not merely a theoretical concept, but provides concrete, engineering-implementable mathematical tools, ensuring high-precision real-time state monitoring even with limited onboard computing resources.

[0071] Taking an eVTOL aircraft performing complex urban air mobility missions as an example. Before takeoff, the system generates theoretical curves based on mission planning. During flight, the BMS collects high-frequency data. If the extended Kalman filter algorithm is used, the system constructs a second-order RC equivalent circuit model, uses the Jacobian matrix to linearize the model equations, updates the SOC and SOE estimates every 5 seconds, and updates the SOP estimate every 200 milliseconds. If the battery is in the low-temperature start-up phase, with strong nonlinearity, the system can switch to or configure the unscented Kalman filter algorithm. By selecting 2n+1 sigma points (n being the dimension of the state variable) for propagation, it more accurately estimates the voltage drop and power capacity of the battery under high-current discharge, avoiding overestimation of SOP due to linearization errors, thereby preventing flight risks caused by insufficient power. Simultaneously, both algorithms support online parameter identification, allowing real-time adjustment of internal resistance and capacitance parameters in the model to adapt to the aging state of the battery.

[0072] Furthermore, this application may select sampling point unscented Kalman filtering or central difference unscented Kalman filtering as the specific implementation of unscented Kalman filtering; on microcontrollers with extremely limited computing resources, a simplified version of extended Kalman filtering may be selected, such as ignoring some cross-coupling terms; for batteries with specific chemical systems, electrochemical models and Kalman filtering algorithms may be combined to form a hybrid estimation algorithm.

[0073] This application employs either the Extended Kalman Filter (EKF) or Unscented Kalman Filter (UKF) algorithm to achieve joint estimation across multiple time scales. This enables the system to effectively handle the strong nonlinear characteristics in the battery state equation, avoiding estimation biases inherent in traditional linear methods under complex conditions. This improves the accuracy and robustness of state of charge (SOC), power state (PS), and energy state (ESW) estimations. By introducing the UKF algorithm, the system avoids the accumulation of linearization errors when the model exhibits high nonlinearity or significant measurement noise, further enhancing the reliability of state estimation under extreme conditions and contributing to flight safety. The use of the EKF algorithm maintains high estimation accuracy while keeping computationally low, facilitating real-time operation in airborne embedded systems and improving the system's engineering practicality and response speed.

[0074] based on Figure 1 The adaptive detection method for aircraft battery status described in this application provides a specific implementation of energy feasibility determination. Its protection boundary lies in evaluating the ability of remaining energy to support subsequent flight missions by comparing the real-time energy status with the theoretically predicted energy status. This is one of the core components of forward-looking mission feasibility determination.

[0075] Please refer to the details. Figure 4 , Figure 4 This is a flowchart illustrating the fourth embodiment of the adaptive detection method for aircraft battery status provided in this application.

[0076] like Figure 4 As shown, the specific steps are as follows: Step S41: Compare the state of charge or energy state in the real-time state estimate with the predicted value at the corresponding moment in the theoretical state evolution curve.

[0077] In the embodiments of this application, the state of charge (SOC) refers to the percentage of the battery's current remaining charge relative to its total capacity, characterizing the battery's charge reserve level; the state of energy (SOE) refers to the total electrical energy that the battery can currently release, which is usually determined by the SOC, discharge rate, temperature, and battery health status, characterizing the battery's actual work capacity; the predicted value at the corresponding moment in the theoretical state evolution curve refers to the theoretical SOC or SOE value that the battery should reach at a specific time point, calculated in the digital twin model based on the initial task planning and the battery model.

[0078] The purpose of comparing real-time state estimates with theoretical predictions is to establish a deviation detection mechanism. Its working principle is that the digital twin model, before flight, has already derived an ideal battery consumption trajectory—the theoretical curve—based on mission planning (such as waypoints and velocity profiles). During flight, the system acquires the actual state of charge (SOC) or state of energy in real time. By calculating the difference between the real-time value and the theoretical value, the degree of deviation between the actual and expected battery performance can be quantified. This comparison is not a simple numerical subtraction, but rather a logical judgment based on preset thresholds. For example, if the real-time SOC is lower than the theoretical prediction by more than a certain percentage, it indicates that the battery is consuming too quickly or its capacity decay is exceeding expectations.

[0079] Step S42: Determine whether the remaining energy supports the completion of subsequent flight missions.

[0080] In this embodiment, the design purpose of determining whether remaining energy supports the completion of subsequent flight missions is to achieve mission feasibility confirmation based on energy reserves. Its working logic is to use the current real-time energy state (or state of charge) as a starting point, and combine it with the energy consumption model of the remaining mission path to calculate the minimum energy requirement needed to complete the remaining mission. Alternatively, it can directly compare the current remaining energy with the energy required to complete the remaining mission. If the real-time energy state is lower than the theoretical energy threshold required to complete the remaining mission, then energy feasibility is determined to be unsatisfactory. This design allows the system to anticipate the risk of power outages or inability to return mid-flight.

[0081] Take an eVTOL aircraft performing long-distance logistics delivery as an example. During mid-flight, the digital twin model predicts a state of charge (SBC) of 60% and an energy state of 55% of the remaining total energy. However, the real-time estimation module detects that due to increased load or lower ambient temperature, the current SBC is only 50% and the energy state is 45%. The system comparison reveals that the real-time values ​​are lower than the predicted values, and further calculations show that the remaining energy is insufficient to support the minimum energy reserves required for return and landing at an alternate airport. At this point, the system determines that energy feasibility is not met and triggers an early warning.

[0082] This application can use the state of energy (SGE) rather than the state of charge (SOC) as the criterion because SGE better reflects the actual usable capacity of the battery under high-rate discharge. For example, in low-temperature environments, the battery's internal resistance increases, and although the SOC may appear sufficient, the usable energy is significantly reduced. Using SGE can more accurately reflect the actual risk, especially when battery aging causes significant changes in the SOC-SOE nonlinear relationship. The theoretical prediction can be calculated based on a more refined thermo-electric coupling model, considering the impact of temperature on energy output. Real-time state estimates can be obtained using an unscented Kalman filter algorithm to improve estimation accuracy.

[0083] This application compares the state of charge or energy state in real-time state estimates with the predicted values ​​at corresponding moments in the theoretical state evolution curve. This allows the system to quantify the deviation between actual and expected battery performance, and promptly identify problems such as rapid energy consumption caused by battery aging, environmental changes, or abnormal loads. By determining whether the remaining energy can support the completion of subsequent flight missions, the system can proactively assess the matching relationship between remaining energy and remaining mission requirements before the risk of insufficient energy occurs. This enables early warning of flight safety risks and avoids safety accidents such as mid-flight power outages or inability to return to base.

[0084] based on Figure 1 The adaptive detection method for aircraft battery status, as described in this application, provides a specific implementation of power feasibility determination. Its protection boundary is defined by clarifying three specific steps: predicting peak power demand, comparing peak power demand with real-time power status, and determining whether peak power output capability meets the demand.

[0085] Please refer to the details. Figure 5 , Figure 5 This is a flowchart illustrating the fifth embodiment of the adaptive detection method for aircraft battery status provided in this application.

[0086] like Figure 5 As shown, the specific steps are as follows: Step S51: Predict the peak power demand for future operating conditions based on the aircraft mission planning information.

[0087] In this embodiment, the peak power requirement for predicting future operating conditions is discussed. Future operating conditions refer to the dynamic states the aircraft will experience in subsequent flight phases, derived from aircraft mission planning information (such as waypoint sequences, altitude profiles, and velocity profiles). Peak power requirement refers to the maximum instantaneous power that the battery system needs to provide under specific operating conditions (such as climb, hover, or maneuver). This step calculates the required torque and speed of the drive motor using the aircraft's kinematic and aerodynamic models, combined with parameters such as the rate of change of flight attitude and rate of change of speed within the current moment and a preset future time window, and then converts this into the battery's peak power requirement. This design allows the system to know in advance the power load the battery needs to withstand, providing a clear comparison target for subsequent feasibility assessments. An understandable embodiment includes: when an eVTOL aircraft performs a logistics delivery mission, the system knows from navigation planning that it will enter the climb phase in 30 seconds. By obtaining the current acceleration command and aircraft mass from flight control data, and using the motor efficiency curve and transmission ratio, the peak power required by the battery during the climb phase is calculated to be 50 kilowatts.

[0088] Step S52: Compare the peak power demand with the power state in the real-time state estimate.

[0089] In this embodiment, the comparison involves the peak power demand and the power state in the real-time state estimate. The power state in the real-time state estimate refers to the maximum power output capability of the battery under current conditions, estimated by the battery management system based on parameters such as the current battery state of charge, temperature, and internal resistance. This comparison step numerically compares the predicted peak power required for future operating conditions with the currently estimated battery power state. This design establishes a direct mapping relationship between "task demand" and "battery supply," enabling the system to assess whether the battery has sufficient power margin to cope with upcoming high-power operating conditions before the actual power demand occurs. An understandable embodiment includes: the system predicts that a peak power of 45 kW is required to perform a maneuver in 20 seconds, while the current real-time estimated battery power state is 40 kW; the system records this difference.

[0090] Step S53: Determine whether the battery's peak power output capability meets future operating conditions.

[0091] In this embodiment, the determination of whether the battery's peak power output capability meets future operating conditions is addressed. This determination is based on a comparison result. If the real-time power state is lower than the peak power demand by more than a preset safety threshold (e.g., 10%), the battery's peak power output capability is determined not to meet future operating conditions, i.e., power feasibility is deemed unsatisfactory. Conversely, if the real-time power state is higher than or equal to the peak power demand plus a safety margin, it is determined to meet the demand. This design transforms abstract state comparisons into concrete feasibility conclusions, directly serving subsequent early warning or strategy adjustment decisions. An understandable embodiment includes: when the predicted peak power demand is 45 kW, and the real-time power state is 40 kW, since 40 kW is lower than 45 kW and the difference exceeds the preset 10% threshold, the system determines that power feasibility is unsatisfactory, thereby triggering a power insufficiency warning and suggesting a reduction in maneuvering amplitude or a delayed ramp-up.

[0092] This application predicts peak power requirements for future operating conditions based on aircraft mission planning information, enabling the system to establish a mission-oriented power demand benchmark. This expands the detection dimension from the current instantaneous state to future operating condition requirements, facilitating early identification of battery overload risks caused by high-power conditions. By comparing peak power requirements with power states estimated in real-time, the system can quantitatively assess the gap between the battery's current power capability and future requirements before the actual power demand occurs, allowing for early intervention to address potential power insufficiency risks. By determining whether the battery's peak power output capability meets future operating condition requirements, the system can make clear feasibility conclusions based on power margins, helping to avoid aircraft loss of control or crashes due to insufficient battery peak power, further improving flight mission safety.

[0093] based on Figure 1 The adaptive detection method for aircraft battery status, as described in this application, provides further limitations on the power feasibility determination step. Its protection boundary lies in clarifying the specific source and constituent operating condition type of the peak power demand. This includes: determining the peak power demand based on a predicted sequence of future operating conditions, and ensuring that the future operating conditions include at least one of climb, cruise, and hovering.

[0094] The future operating condition sequence refers to the sequence of flight states that the aircraft will experience over a period of time, derived from the aircraft's current navigation planning information (such as waypoints, altitude profiles, and velocity profiles); peak power demand refers to the maximum instantaneous power output of the battery required by the aircraft's power system to maintain that state or complete a maneuver under a specific operating condition; climb operating condition refers to the flight phase in which the aircraft increases its altitude vertically or obliquely; cruise operating condition refers to the phase in which the aircraft flies horizontally at a predetermined altitude and speed; hover operating condition refers to the flight phase in which the aircraft maintains a relatively stationary position in the air.

[0095] The design principle behind determining peak power requirements based on predicted future operating conditions is that the power challenges faced by batteries during flight do not arise from the current moment, but from impending dramatic maneuvers. For example, a sudden switch from a stable cruise state to a high-speed climb requires the motors to draw a massive current instantaneously, causing a sharp drop in battery voltage. This impact cannot be predicted by only monitoring the current power state. By analyzing the waypoint sequence and kinematic parameters (such as acceleration and rate of altitude change) in the mission plan, combined with the aircraft's dynamics model, the theoretical power required at each future moment can be calculated. Taking the maximum of these power demands as the peak power requirement allows for accurate quantification of the power capacity the battery needs to store. This is analogous to calculating the torque needed for a hill climb before driving, rather than reacting only when the wheels slip.

[0096] The future operating conditions include at least climb, cruise, and hovering. This is because these three conditions represent the most typical operating modes for eVTOL and drones, with vastly different battery requirements. Climbing typically has the highest peak power demand due to the need to overcome gravity and high acceleration; hovering, while potentially having lower peak power, offers continuous and stable power output, placing high demands on battery thermal management and sustained discharge capabilities; cruise offers relatively stable power but lasts longer, impacting energy consumption. Including these three conditions covers the main power characteristics of the aircraft throughout its entire mission cycle. Typical mission profiles of multi-rotor aircraft can be selected, with takeoff / hovering, accelerated climb, and level flight / cruise being key nodes. For example, in logistics delivery missions, the power surge from takeoff to hovering, followed by accelerated climb to 200 meters, is the most dramatic and the primary source of peak power demand. If the mission includes cargo delivery, power stability during hovering is equally crucial. Clearly defining these conditions makes power feasibility assessments more specific and comprehensive, allowing for targeted evaluation of battery performance under different typical loads.

[0097] Take an eVTOL (eVehicle-to-Air Vehicle) system performing urban aerial logistics missions as an example. Before takeoff, the digital twin model analyzes the mission plan and determines that the aircraft will hover for takeoff within 0-10 seconds, climb to 200 meters within 10-30 seconds, and cruise at 20 m / s within 30-120 seconds. Based on the dynamic equations, the model calculates that the hovering phase requires 20 kW of power, the climb phase requires overcoming gravity and acceleration, resulting in an instantaneous peak power demand of 45 kW, and the cruise phase requires 15 kW. Therefore, the system determines the peak power demand within the next 30 seconds to be 45 kW (from the climb condition). During flight, when the aircraft is in the cruise phase (current power 15 kW), the system, based on the predicted future sequence of conditions, identifies in advance that it will enter the climb condition in 30 seconds, with a peak demand of 45 kW. At this time, the system checks whether the current battery SOP (Start of Operation) is greater than 45 kW. If the SOP is only 40 kW, it determines that the power feasibility is not met, thus issuing an early warning. It can be replaced by other models, such as fixed-wing UAVs, whose future operating conditions may include takeoff, climb, cruise, approach and landing. Takeoff and climb are the main sources of peak power, and the same logic applies.

[0098] This application can use more complex dynamic models, such as nonlinear models that consider the change of air drag coefficient with speed, to calculate power demand; future operating conditions may include high dynamic conditions such as crosswind compensation and emergency obstacle avoidance maneuvers; the determination of peak power demand can be based on statistical methods, taking the 95th percentile of the power distribution of future operating conditions instead of the maximum value, in order to balance safety margin and system cost.

[0099] This application determines peak power demand based on a predicted sequence of future operating conditions, enabling the system to quantify impending power surges in advance based on the aircraft's kinematic characteristics and mission planning. This extends the power detection dimension from the current moment to future time periods, allowing for early detection of potential power risks. By specifically defining future operating conditions as one or more of climb, cruise, and hovering, the power feasibility assessment can cover the diverse battery power output requirements of the aircraft in different typical flight phases. This facilitates a comprehensive evaluation of the battery's power adaptability in complex flight missions, ensuring sufficient capacity margin during the climb phase with the highest power demand or the hovering phase with sustained high load.

[0100] based on Figure 1 The aforementioned adaptive detection method for aircraft battery status provides a closed-loop parameter correction mechanism for a digital twin model. Its protection boundary lies in clearly defining the data type of feedback (voltage, current, temperature, SOC) and the specific physical model parameters to be corrected (internal resistance, capacity decay coefficient), thereby achieving dynamic adaptation to battery aging characteristics.

[0101] Please refer to the details. Figure 6 , Figure 6 This is a flowchart illustrating the sixth embodiment of the adaptive detection method for aircraft battery status provided in this application.

[0102] like Figure 6 As shown, the specific steps are as follows: Step S61: During flight or when the judgment result does not meet the preset conditions, the actual collected voltage, current, temperature, and SOC data are fed back to the digital twin model.

[0103] In this embodiment of the application, the digital twin model refers to a virtual mapping entity that includes an electrical model, a thermal model, and a capacity aging model.

[0104] Determining the necessity of feedback data is crucial. Voltage, current, and temperature are direct physical characteristics of the battery's electrochemical state, while SOC (State of Charge) is a state index estimated by combining these physical quantities. During flight, the actual output characteristics of the battery often deviate from the theoretical model. This deviation mainly stems from battery manufacturing tolerances, micro-damage during use, and environmental stress. By collecting these four types of data in real time, we can provide the model with a true "ground truth," which can be used to calibrate the model's input-output mapping relationship.

[0105] Step S62: By correcting the model parameters, the subsequent deductions are made more accurate. The model parameters include at least the internal resistance value and the capacity decay coefficient.

[0106] In the embodiments of this application, the internal resistance value refers to the battery's ability to impede the flow of current, which varies with SOC, temperature and cycle number; the capacity decay coefficient refers to the ratio of the battery's current usable capacity to its nominal capacity, used to quantify the degree of degradation of the battery's state of health (SOH).

[0107] Explain the logic behind model parameter correction. Battery aging is not a linear process, but rather a nonlinear process influenced by operating conditions (such as high-rate charging and discharging, and high temperatures). An increase in internal resistance leads to a larger voltage drop, thus affecting the accuracy of the voltage response model; changes in the capacity decay coefficient directly determine the slope and endpoint of the energy state evolution curve. By comparing the residuals between real-time collected voltage and current data and model predictions, algorithms such as least squares or recursive least squares can be used to identify and update the internal resistance online. For example, if the measured voltage is lower than the predicted voltage while the current is the same, it can be inferred that the internal resistance has increased. Simultaneously, by monitoring the cumulative change in SOC over a long period and the deviation from the theoretical prediction, the capacity decay coefficient can be corrected. This design makes the model no longer a static mathematical formula, but a dynamic system with "memory" and "learning ability."

[0108] Take an eVTOL aircraft performing multiple logistics delivery missions as an example. After the first flight, the system found that the measured State of Charge (SOC) decreased 2% faster than the model predicted. By reviewing the current integral data during the flight, the system determined that the battery capacity had slightly decreased, and therefore adjusted the capacity decay factor from 1.0 to 0.98. In the second flight, the model used the corrected capacity factor to re-derive the theoretical curve. Simultaneously, if, during a high-power ramp-up, the measured voltage drop was greater than the model prediction, the system detected an abnormal increase in internal resistance and updated the internal resistance parameter. An online parameter identification algorithm based on extended Kalman filtering can be used to fine-tune the internal resistance and capacity factor every second of flight, ensuring that the model always closely reflects the actual physical state of the battery.

[0109] This application utilizes actual collected voltage, current, temperature, and SOC data to enable the digital twin model to obtain a benchmark for the battery's physical response under real flight conditions, providing reliable data support for parameter correction. By correcting the internal resistance value, the voltage response model can accurately reflect the dynamic characteristics of the battery's internal impedance changing with aging and operating conditions, eliminating voltage prediction errors caused by internal resistance drift, thereby improving the accuracy of energy feasibility assessment. By correcting the capacity decay coefficient, the capacity evolution model can adapt to the personalized aging trajectories of batteries from different aircraft due to differences in usage frequency and operating conditions, eliminating remaining energy prediction errors caused by biases in battery health state estimation, thus achieving real-time alignment between the theoretical state evolution curve and the actual physical state of the battery, further improving the forward-looking prediction accuracy of subsequent mission simulations.

[0110] based on Figure 1 The aforementioned adaptive detection method for aircraft battery status provides specific content of the generated flight strategy adjustment instructions. Its protection boundary is to clarify the specific flight control measures that the system can take when the battery status does not meet the mission feasibility requirements.

[0111] Reducing cruise speed refers to decreasing the target airspeed of the aircraft during the level flight phase; reducing climb rate refers to decreasing the rate of ascent in the vertical direction; modifying waypoint sequence refers to replanning the time or spatial order of subsequent waypoints; returning to base refers to changing the flight destination to the takeoff point or alternate landing point.

[0112] When a feasibility assessment for a forward-looking mission indicates insufficient energy availability, reducing cruise speed is employed. The design principle is that an aircraft's energy consumption is closely related to its speed; reducing speed decreases energy consumption per unit time, thereby extending the flight time supported by remaining energy, allowing the aircraft to complete the remaining distance or safely reach an alternate landing point. For example, with a logistics drone, when it detects that the remaining battery power is insufficient to support the planned return flight, the system automatically reduces the cruise speed from 25 meters per second to 18 meters per second to reduce power requirements and ensure it can fly back to base.

[0113] When the displayed power is deemed insufficient, meaning the battery cannot provide the peak power required for the upcoming operating condition, a reduced climb rate is employed. The design rationale is that climb conditions typically require maximum instantaneous power output; reducing the climb rate directly lowers the instantaneous torque and current required by the motors, thus preventing the battery from cutting off output or experiencing a sudden voltage drop due to overload protection. For example, in an eVTOL aircraft, if a climb to 2000 meters is predicted, but the current operating point (SOP) is insufficient to support the planned rapid climb, the system instructs a reduction in the climb rate, adopting a gentler climb path to match the battery's real-time power output capability.

[0114] Modifying waypoint sequence is employed to optimize energy or power distribution. The design logic involves adjusting the mission sequence to postpone high-energy-consuming or high-power-demanding conditions until when the battery is in better condition, or skipping certain non-critical high-energy-consuming waypoints. For example, the waypoint sequence could be adjusted to first execute low-energy-consuming cruise missions, and then execute high-power hovering missions after the battery has partially recovered voltage stability through some means (such as taxiing). Alternatively, complex terrain requiring high-power takeoffs and landings could be changed to flat ground takeoffs and landings.

[0115] Returning to home is employed as the safest fallback strategy. Its design purpose is to force the aircraft to return to the safest known area when energy or power is severely insufficient, and adjusting speed or waypoints cannot guarantee safe mission completion. For example, when both energy and power feasibility criteria are not met, and the destination is too far away, the system directly generates a return-to-home command, shuts down non-core payloads, and returns to the takeoff point or the nearest alternate landing point as quickly as possible.

[0116] Take an eVTOL aircraft performing urban air mobility missions as an example. During flight, the digital twin model predicts a strong headwind ahead, leading to a significant increase in estimated energy consumption and failing to meet energy feasibility criteria. In this case, the system generates flight strategy adjustment instructions, including reducing cruise speed to decrease wind resistance and power consumption, and modifying the waypoint sequence to skip two secondary observation points and fly directly to the final destination to ensure sufficient energy to complete the primary mission. Simultaneously, if the system detects an increase in battery temperature leading to a decrease in power capability, it also instructs a reduction in the climb rate to avoid high-power maneuvers when battery condition is poor. If these adjustments still cannot guarantee safety, a return-to-home instruction is triggered, planning the shortest and least energy-intensive return route.

[0117] Furthermore, this application can choose to reduce the motor speed instead of reducing the cruise speed, with equivalent effects; it can choose to increase the flight altitude to utilize airflow gliding instead of reducing the climb rate, provided that energy allows; it can choose to add temporary waypoints to avoid high-energy-consuming areas instead of modifying the waypoint order; and it can choose automatic landing points instead of return-to-home points, which is suitable for emergency landing scenarios.

[0118] This application, by limiting flight strategy adjustment commands including reducing cruise speed, enables the system to proactively reduce the aircraft's energy consumption rate when energy feasibility is insufficient, thereby extending the remaining flight time and facilitating the completion of the designated mission or safe arrival at the alternate landing point. By limiting flight strategy adjustment commands including reducing the rate of climb, the system can reduce instantaneous power demand when power feasibility is insufficient, preventing battery voltage collapse or protection shutdown due to overload, thus ensuring power stability during flight. By limiting flight strategy adjustment commands including modifying waypoint sequence, the system can flexibly adjust the mission execution sequence based on the real-time energy and power status of the battery, optimizing high-energy consumption conditions to moments when the battery state is optimal, thus facilitating dynamic matching between flight missions and battery capabilities. By limiting flight strategy adjustment commands including returning to base, the system can immediately implement the highest level of safety measures when both energy and power are severely insufficient and cannot be recovered through other strategies, thus maximizing the safety of the aircraft and personnel.

[0119] Those skilled in the art will understand that, in the above-described method of the specific implementation, the order in which each step is written does not imply a strict execution order and does not constitute any limitation on the implementation process. The specific execution order of each step should be determined by its function and possible internal logic.

[0120] To implement the aforementioned adaptive detection method for aircraft battery status, this application also proposes an adaptive detection system for aircraft battery status, for details please refer to [link / reference needed]. Figure 7 , Figure 7This is a schematic diagram of an embodiment of the adaptive detection system for aircraft battery status provided in this application.

[0121] The adaptive detection system 500 for aircraft battery status in this embodiment includes: The digital twin module 51 is used to acquire the mission planning information of the aircraft, use a pre-built battery-aircraft digital twin model to virtually simulate the flight mission, and generate theoretical state evolution curves for each stage of flight. The theoretical state evolution curves include at least the state of charge evolution curve, the power state evolution curve, and the energy state evolution curve.

[0122] The state estimation module 52 is used to collect flight state data and battery parameter data in real time during flight, and to perform joint estimation of the battery's state of charge, state of power and state of energy at multiple time scales to obtain real-time state estimates.

[0123] The feasibility assessment module 53 is used to compare the theoretical state evolution curve with the real-time state estimate to make a forward-looking task feasibility assessment, which includes energy feasibility assessment and power feasibility assessment.

[0124] The adaptive adjustment module 54 is used to generate an early warning signal and / or flight strategy adjustment command when the result of the feasibility determination of the prospective mission does not meet the preset conditions, and to feed the real-time data back to the digital twin module to update the theoretical state evolution curve.

[0125] To implement the aforementioned adaptive detection method for aircraft battery status, this application also proposes an aircraft, for details please refer to [link / reference needed]. Figure 8 , Figure 8 This is a structural schematic diagram of an embodiment of the aircraft provided in this application.

[0126] The aircraft 400 in this embodiment includes the aforementioned aircraft battery state adaptive detection system 41.

[0127] The above description is merely an embodiment of this application and does not limit the patent scope of this application. Any equivalent structural or procedural transformations made using the content of this application's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of this application.

Claims

1. An adaptive detection method for aircraft battery status, characterized in that, Includes the following steps: The mission planning information of the aircraft is obtained, and the flight mission is virtually simulated using a pre-built battery-aircraft digital twin model to generate theoretical state evolution curves for each stage of flight. The theoretical state evolution curves include at least the state of charge evolution curve, the power state evolution curve, and the energy state evolution curve. During flight, flight status data and battery parameter data are collected in real time, and the battery's state of charge, state of power, and state of energy are jointly estimated at multiple time scales to obtain real-time state estimates. The theoretical state evolution curve is compared with the real-time state estimate to make a forward-looking task feasibility determination, which includes energy feasibility determination and power feasibility determination. If the result of the feasibility assessment of the prospective mission does not meet the preset conditions, an early warning signal and / or flight strategy adjustment instruction are generated, and the real-time data is fed back to the digital twin model to update the theoretical state evolution curve. The flight strategy adjustment instructions include one or more of the following: reducing cruise speed, reducing climb rate, modifying waypoint order, or returning to base.

2. The method according to claim 1, characterized in that, The battery-aircraft digital twin model includes an electrical model, a thermal model, and a capacity model; The electrical model is used to simulate the voltage response and power output characteristics of a battery under different operating conditions. The thermal model is used to simulate the temperature rise characteristics of the battery during flight; The capacity model is used to simulate the capacity decay characteristics of a battery.

3. The method according to claim 1, characterized in that, The virtual simulation includes: Based on the waypoint sequence, altitude profile, and velocity profile in the mission planning information, predict the operating condition sequence for each stage of flight; The theoretical state evolution curve is generated by simulating the evolution of battery state parameters based on the operating condition sequence.

4. The method according to claim 1, characterized in that, The joint estimation across multiple time scales includes: The state of charge and the state of energy are estimated at a first time scale, which is 5-10 seconds. The power state is estimated on a second time scale, which is 100-500 milliseconds; Wherein, the first time scale is larger than the second time scale.

5. The method according to claim 1, characterized in that, The joint estimation of the battery's state of charge, state of power, and state of energy across multiple time scales is achieved using either the extended Kalman filter algorithm or the unscented Kalman filter algorithm.

6. The method according to claim 1, characterized in that, The energy feasibility assessment includes: Compare the state of charge or energy state in the real-time state estimate with the predicted value at the corresponding moment in the theoretical state evolution curve; Determine whether the remaining energy is sufficient to complete the subsequent flight mission.

7. The method according to claim 1, characterized in that, The power feasibility assessment includes: Predict the peak power requirements for future operating conditions based on the aforementioned aircraft mission planning information; Compare the peak power demand with the power state in the real-time state estimate; Determine whether the battery's peak power output capability meets future operating conditions. The peak power demand is determined based on a predicted sequence of future operating conditions, which include at least one or more of the following: climbing, cruising, and hovering conditions.

8. The method according to claim 1, characterized in that, The step of feeding the real-time data back to the digital twin model to update the theoretical state evolution curve specifically includes: During flight or when the judgment result does not meet the preset conditions, the actual collected voltage, current, temperature, and SOC data will be fed back to the digital twin model. By correcting the model parameters, subsequent calculations can be made more accurate. The model parameters include at least the internal resistance value and the capacity decay coefficient.

9. An adaptive detection system for aircraft battery status, characterized in that, include: The digital twin module is used to acquire mission planning information of the aircraft, perform virtual simulation of the flight mission using a pre-built battery-aircraft digital twin model, and generate theoretical state evolution curves for each stage of flight. The theoretical state evolution curves include at least the state of charge evolution curve, the power state evolution curve, and the energy state evolution curve. The state estimation module is used to collect flight state data and battery parameter data in real time during flight, and to perform joint estimation of the battery's state of charge, state of power and state of energy at multiple time scales to obtain real-time state estimates. The feasibility assessment module is used to compare the theoretical state evolution curve with the real-time state estimate to make a forward-looking task feasibility assessment, which includes energy feasibility assessment and power feasibility assessment. An adaptive adjustment module is used to generate an early warning signal and / or flight strategy adjustment command when the result of the feasibility determination of the prospective mission does not meet the preset conditions, and to feed the real-time data back to the digital twin module to update the theoretical state evolution curve.

10. An aircraft, characterized in that, Includes the adaptive detection system for aircraft battery status as described in claim 9.

Images