A method and computing device for determining aircraft power failure

Abstract

The embodiments of this specification provide a method and computing device for determining aircraft power failure, which is performed using a speed prediction model. The speed prediction model is a machine learning model, which obtains the initial speed, motor status and throttle command input to the target motor in the target aircraft as input signals. The speed prediction model is used to determine a first predicted speed of the target motor based on the input signal, and the fault detection result of the target motor is determined based on the difference between the first predicted speed and the actual speed of the target motor after receiving the throttle command. On the one hand, it can be determined in real time and accurately whether any motor in the aircraft has a fault; on the other hand, the first predicted speed output by the speed prediction model provides an objective and reliable standard for judging aircraft power failure.

Description

Technical Field

[0001] The embodiments of this specification belong to the field of data processing technology, and more specifically, to a method and computing device for determining aircraft power failure. Background Art

[0002] Multi-rotor EVTO aircraft play an important role in many fields, including aerial photography, plant protection, inspection, patrol, logistics, and passenger transport. However, as their application becomes increasingly widespread, accidents such as loss of control and crashes become more common. These accidents not only cause property damage but can also threaten human life. Therefore, higher requirements are being placed on the flight safety and reliability of multi-rotor EVTO aircraft.

[0003] Currently, methods for detecting power failures in aircraft still have deficiencies in terms of accuracy and real-time performance, making it difficult to accurately determine the rotor or specific component where the fault is located immediately after a failure occurs. For example, traditional high-precision nonlinear dynamic models can describe the ideal flight state of a multi-rotor aircraft, but high-precision nonlinear flight dynamics models require high hardware computing power, and conventional airborne computing equipment cannot meet the real-time requirements of their calculations. Alternatively, empirical data can be used to preset the abnormal parameter value ranges corresponding to the various parameters collected by each sensor. However, due to the difference between empirical data and the actual flight state, using such abnormal parameter value ranges is also difficult to detect the aircraft's power failure immediately.

[0004] Thus, the present invention provides a method and computing device for determining aircraft power failure. Summary of the Invention

[0005] The embodiments of this specification aim to provide a method and computing device for determining aircraft power failure.

[0006] In one aspect, the present specification provides a method for determining an aircraft power failure, wherein the method is performed using a rotational speed prediction model, wherein the rotational speed prediction model is a machine learning model, and the method includes:

[0007] Acquire an initial speed and a motor state of a target motor in the target aircraft, and a throttle command input to the target motor as input signals;

[0008] Determining a first predicted speed of the target motor according to the input signal using the speed prediction model;

[0009] A fault detection result of the target motor is determined according to a difference between the first predicted speed and an actual speed of the target motor after receiving the throttle command.

[0010] In some implementations, the motor state specifically includes at least one of a motor voltage, a motor current, and a rotor torque load.

[0011] In some implementations, determining a fault detection result of the target motor based on a difference between the first predicted speed and an actual speed of the target motor after receiving the throttle command specifically includes:

[0012] determining a current speed difference according to a difference between the first predicted speed and an actual speed of the target motor after receiving the throttle command;

[0013] determining a speed loss ratio according to a ratio of the current speed difference to the first predicted speed;

[0014] A fault detection result of the target motor is determined according to the speed loss ratio and a preset fault threshold.

[0015] In some implementations, determining a fault detection result of the target motor based on the speed loss ratio and a preset fault threshold specifically includes:

[0016] When the speed loss ratio is greater than a preset fault threshold, determining that the target motor is a faulty motor;

[0017] When the target motor is a faulty motor, determining an expected pulling force of the target motor according to the first predicted speed, and determining an actual pulling force according to the actual speed;

[0018] According to the difference between the actual pulling force and the expected pulling force, a power loss rate corresponding to the target motor is determined as a fault detection result corresponding to the target motor.

[0019] In some implementations, after determining the power loss rate corresponding to the target motor, the method further includes:

[0020] A fault-tolerant control strategy for the target aircraft is determined according to the power loss rate.

[0021] In some implementations, after determining that the target motor is a faulty motor, the method further includes:

[0022] A fault detection result of the target motor is determined according to the motor state and the speed loss ratio.

[0023] In some implementations, the method further includes:

[0024] Acquire a training sample, wherein the training sample includes an input sample and a sample label, wherein the input sample includes a motor state, an initial speed, and a throttle command of a sample motor, and the sample label includes a future speed corresponding to the input sample;

[0025] Inputting the input sample into the speed prediction model to be trained, and determining a second predicted speed output by the speed prediction model to be trained;

[0026] The speed prediction model to be trained is trained according to the difference between the second predicted speed and the future speed.

[0027] In some implementations, the method further includes:

[0028] Establish a simulation dynamics model corresponding to the target aircraft according to the control instruction module, motor and electric control model, aerodynamic model, dynamic equations, and kinematic equations corresponding to the target aircraft;

[0029] Using the simulation dynamics model, simulating the motor speed, motor state, and throttle command input to each motor of the target aircraft under various preset operating conditions;

[0030] The training samples corresponding to the target aircraft are determined according to the simulation results.

[0031] In some implementations, the rotational speed prediction model is deployed on an onboard computing device of the target aircraft.

[0032] A second aspect of this specification provides a computing device, including a memory and a processor, wherein the memory stores executable code, and when the processor executes the executable code, the method described in the first aspect is implemented.

[0033] The solution for determining aircraft power failure provided in the embodiments of this specification can, on the one hand, accurately determine in real time whether any motor in the aircraft has failed; on the other hand, the first predicted speed output by the speed prediction model provides an objective and reliable standard for judging aircraft power failure. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] In order to more clearly illustrate the technical solutions of the embodiments of this specification, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments recorded in this specification. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0035] Figure 1 This is a schematic structural diagram of an aircraft power failure determination system provided by an embodiment of the present invention;

[0036] Figure 2 This is a flow chart of a method for determining an aircraft power failure provided by an embodiment of the present invention;

[0037] Figure 3 It is a flowchart of a method for training a speed prediction model provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0038] To help those skilled in the art better understand the technical solutions in this specification, the following will provide a clear and complete description of the technical solutions in the embodiments of this specification, in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of this specification, not all of them. All other embodiments derived by those skilled in the art based on the embodiments in this specification without creative effort shall fall within the scope of protection of this specification.

[0039] The power failure of an aircraft may be due to a variety of reasons, such as mechanical failure, electric control failure, circuit failure, etc. Generally, the power failure of an aircraft is always manifested as a stall of the aircraft's motor. A multi-rotor EVTOL aircraft may have several motors that provide power to different rotors, and each motor works together to maintain the flight state of the aircraft. In the prior art, each motor in the same aircraft can be monitored, and when the speed of any motor is significantly lower than that of other motors, it can be determined that the motor has failed. However, this method makes it difficult to locate the faulty motor at the first time the fault occurs, and thus it is impossible to enable the corresponding fault-tolerant control strategy in the first time to reduce the losses caused by the aircraft's power failure.

[0040] thus, Figure 1 A schematic diagram of the structure of an aircraft power failure determination system in accordance with one embodiment of this specification is shown. The aircraft power failure determination system may include a speed prediction model and a fault determination module. The aircraft power failure determination system may be deployed on an aircraft's onboard computing device. The system receives motor status and speed transmitted by onboard sensors, as well as throttle commands for each motor output by each controller. The system then uses the speed prediction model and the fault determination module to determine a fault detection result for each motor.

[0041] The speed prediction model can determine a first predicted speed for each motor based on the throttle command output by the controller for each motor and the motor status of each motor collected by the sensor. The fault determination module can determine the fault detection result for each motor in the aircraft based on the first predicted speed for each motor and the actual speed of each motor after receiving the throttle command. It should be noted that for any motor, the aforementioned fault detection result may only include whether the motor is a faulty motor; if the motor is a faulty motor, it may further include the fault type and fault severity of the motor, etc. This specification does not impose any restrictions on this.

[0042] In some implementations, the speed prediction model may be a neural network model, such as a deep neural network (DNN), a residual network (ResNet), etc.

[0043] Figure 2 A flow chart of a method for determining an aircraft power failure in accordance with an embodiment of this specification is shown. This method utilizes a rotational speed prediction model, which is a machine learning model. This rotational speed prediction model can be installed in an onboard computing device of a target aircraft or in a remote control terminal of the target aircraft. This specification does not limit this. The method includes:

[0044] S201: Acquire an initial rotation speed, a motor state, and a throttle command input to a target motor in a target aircraft as input signals.

[0045] For the target aircraft, use Figure 2 The method shown in FIG. 1 is to use each motor of the target aircraft as a target motor to determine whether each motor in the target aircraft has a fault, and then determine whether the target aircraft has a power fault. The following only takes the target motor as an example. Figure 2 A method for determining aircraft power failure is introduced as shown.

[0046] First, the initial speed and motor state of the target motor in the target aircraft are acquired through the onboard sensor of the target aircraft, and the throttle command outputted by the controller to the target motor is acquired.

[0047] It should be noted that the throttle command, motor state and initial speed obtained in step S201 correspond to the same control cycle, that is, the target motor will receive the throttle command at the initial speed under the motor state and adjust the output power accordingly according to the throttle command.

[0048] S203: Determine a first predicted speed of the target motor according to the input signal using the speed prediction model.

[0049] The initial speed, motor state and throttle command corresponding to the target motor are used as input signals, and the speed prediction model in the aircraft power failure determination system can be used to determine the first predicted speed of the target motor corresponding to the input signal.

[0050] Specifically, the first predicted speed may refer to the speed of the target motor without faults in the motor state, where the motor speed is the initial speed, after the motor receives the throttle command.

[0051] In some implementations, when training the speed prediction model, any motor of the target aircraft may be prepared in advance as a sample motor, and any control cycle of the sample motor, for example, the control cycle The motor state, initial speed and the throttle command of the sample motor in the motor state are used as sample inputs, and the sample motor is determined after receiving the throttle command - that is, the next control cycle of the control cycle The motor speed is used as the sample label.

[0052] Thus, when the target aircraft operates normally as a whole and the sample motor has no faults, the first predicted speed output by the trained speed prediction model can represent the ideal speed of a motor of the same model as the sample motor, that is, the target motor, after receiving the throttle command.

[0053] The aforementioned normal overall operation of the target aircraft may refer to the absence of any anomalies in the aircraft's software and hardware, and the absence of any external environmental disturbances. Furthermore, if the target motor and the sample motor are of the same model, and the target motor's actual speed after receiving the throttle command differs from the first predicted speed, the target motor may be deemed to have failed.

[0054] In some implementations, when the models of the motors in the target aircraft are the same, the motor states of the motors and the throttle commands corresponding to each motor state can be used to jointly train a same speed prediction model, and the trained speed prediction model can be used to predict the first predicted speed of each motor in the target aircraft.

[0055] S205 : Determine a fault detection result of the target motor according to a difference between the first predicted speed and an actual speed of the target motor after receiving the throttle command.

[0056] After determining the first predicted speed of the target motor, the fault determination module in the aircraft power fault determination system can determine the fault detection result of the target motor based on the difference between the first predicted speed and the actual speed of the target motor after receiving the throttle command.

[0057] It should be noted that, in order to ensure the timeliness of the fault detection result, usually, within a control cycle, for example, the control cycle The controller in the aircraft outputs the control cycle Input the throttle command of the target motor, and the onboard sensor in the aircraft determines the control cycle The motor state and initial speed at the beginning, then you can execute Figure 2 The method shown, according to which the control cycle At the beginning, the motor state, initial speed and throttle command of the target motor are input. After the target motor receives the throttle command, the control cycle is The first predicted speed of the target motor at the beginning. On the other hand, when the target motor receives the throttle command, that is, the control cycle At the beginning, the control cycle determined by the onboard sensors can be obtained The actual speed of the target motor. Thus, the difference between the first predicted speed and the actual speed can be determined within a control cycle (usually, the length of the aircraft control cycle is between 10-100 milliseconds). Further, the following operations are performed in parallel on each motor of the aircraft: Figure 2 The method shown can determine whether the aircraft has a power failure.

[0058] Specifically, as described above, the fault detection result may only include whether the target motor is a faulty motor. In this case, a fault threshold may be preset, which may be expressed as a percentage. The fault determination module may determine the absolute value of the difference between the first predicted speed and the actual speed as the current speed difference, and determine the ratio of the current speed difference to the first predicted speed as the speed loss ratio. The speed loss ratio is then compared with the fault threshold. If the ratio is greater than the fault threshold, the target motor may be determined to be a faulty motor.

[0059] It should be noted that if Figure 2 In the illustrated method, the current speed difference is the difference between the first predicted speed output by the speed prediction model and the actual speed, that is, the difference between the target motor's ideal speed and its actual speed at the same point in time. Unlike the conventional method of determining the speed difference based on the actual speeds acquired at two consecutive time points, the current speed difference obtained in this embodiment more intuitively reflects the likelihood and severity of a target motor failure through the numerical magnitude of the difference.

[0060] Further, according to the introduction corresponding to step S205, the fault threshold in the embodiment of this specification is the threshold corresponding to the speed loss ratio, that is, the fault threshold can be expressed in the form of a percentage, which is different from the abnormal parameter value range directly set for the parameter itself in the prior art.

[0061] The fault threshold can be set by technicians, for example, 5%, 10%, etc. Specifically, a smaller fault threshold indicates greater sensitivity to target motor faults, while a larger fault threshold indicates greater tolerance to random fluctuations caused by external interference. This allows for a certain degree of tolerance for fluctuations in the target motor, avoiding overly sensitive fault detection results.

[0062] like Figure 2The method shown in the figure can, on the one hand, determine in real time and accurately whether any motor in the aircraft has failed; on the other hand, the first predicted speed output by the speed prediction model provides an objective and reliable standard for judging the aircraft power failure.

[0063] In some implementations, the motor state specifically includes at least one of a motor voltage, a motor current, and a rotor torque load.

[0064] In some implementations, Figure 2 In step S205 shown, the current speed difference is determined based on the difference between the first predicted speed and the actual speed of the target motor after receiving the throttle command, the speed loss ratio is determined based on the ratio of the current speed difference to the first predicted speed, and the fault detection result of the target motor is determined based on the speed loss ratio and the preset fault threshold.

[0065] For details, please refer to the above description of step S205, which will not be elaborated in this specification.

[0066] Furthermore, in some implementations, Figure 2 In step S205 shown, when the speed loss ratio is greater than a preset fault threshold, the target motor is determined to be a faulty motor. When the target motor is a faulty motor, the expected pulling force of the target motor is determined based on the first predicted speed, and the actual pulling force is determined based on the actual speed. Based on the difference between the actual pulling force and the expected pulling force, the power loss rate corresponding to the target motor is determined as the fault detection result corresponding to the target motor.

[0067] As previously mentioned, when the target motor is a faulty motor, the fault detection result may further include the fault severity of the target motor. Specifically, a speed-to-tension conversion formula corresponding to the target motor can be determined in advance based on the structural parameters of the target aircraft and the motor control model of the target motor. In step S205, based on this speed-to-tension conversion formula, the expected tension corresponding to the first predicted speed and the actual tension corresponding to the actual speed are determined. Based on the difference between the actual tension and the expected tension, the power loss rate corresponding to the target motor—i.e., the fault severity—is determined as the fault detection result corresponding to the target motor.

[0068] After determining the power loss rate, the additional pulling force required by each motor in the target aircraft to compensate for the power loss of the target motor can be determined based on the power loss rate, and the next control cycle can be further determined based on each additional pulling force. The throttle commands corresponding to the motors that have not failed are used as the fault-tolerant control strategy for the target aircraft.

[0069] The power loss rate determined based on the first predicted speed and the actual speed can accurately describe the power loss of the target motor, thereby providing a more accurate basis for the fault-tolerant control strategy.

[0070] On the other hand, the fault detection result may also include the fault type of the target motor. In some implementations, Figure 2 In step S205 shown, a fault detection result of the target motor is determined according to the motor state and the speed loss ratio.

[0071] The fault determination module can also determine the fault type of the target motor based on the motor state and speed loss ratio of the target motor. Specifically, the technician can preset different fault types, such as mechanical fault, electrical control fault, circuit fault, etc., and further preset the template parameter range corresponding to each fault type. Thus, the fault determination module can compare the motor state and speed loss ratio of the target motor with the preset template parameter ranges. When the motor state and speed loss ratio of the target motor meet any template parameter range, it can be determined that the fault of the target motor is of that fault type.

[0072] For example, the voltage and current of a motor with an electronic control failure will not return to zero, but the speed loss ratio will be high; the current of a motor with a circuit failure will return to zero... Based on this, the template parameter range corresponding to each fault type can be determined.

[0073] On the other hand, when using Figure 2 When the method shown in the figure monitors whether any aircraft has a power failure, generally, each motor of the aircraft needs to be executed once in each control cycle. Figure 2 Thus, for the current control cycle , current control cycle The corresponding speed loss ratio can be obtained for each previous control cycle.

[0074] Furthermore, the fault detection results may also include dynamic detection results. Specifically, the dynamic detection results can indicate the changing trend of speed loss, such as a momentary fault or a gradual power loss. Furthermore, these dynamic detection results can provide a basis for possible subsequent maintenance and adjustments of the target motor.

[0075] In some implementations, after determining the current speed difference, the speed loss ratio corresponding to the current control cycle is determined based on the ratio of the current speed difference to the first predicted speed, and the dynamic detection result of the target motor is determined based on the speed loss ratio corresponding to the current control cycle and the speed loss ratio corresponding to each historical control cycle.

[0076] Specifically, the dynamic comparison length n can be preset to set the current control period The previous n consecutive control cycles—— As time series data formed based on the speed loss ratios of each historical control cycle and the current control cycle, a speed loss change trend corresponding to the current control cycle is determined as the dynamic detection result of the target motor.

[0077] It should be noted that because the speed prediction model's inference process does not require the extensive nonlinear computations required by the nonlinear dynamics equations, its inference speed can match the length of each control cycle. This ensures that the target motor's dynamic detection results are determined no longer than the control cycle, thus maintaining the timeliness of the dynamic detection results. If the prior art aircraft's nonlinear dynamics equations were used to directly calculate the target motor's first predicted speed for each control cycle, the calculation would take longer than the length of a single control cycle. Calculating the speed loss change rate based on this calculation would further reduce the timeliness of the dynamic detection results due to the added delay.

[0078] Figure 3 A flow chart of a method for training a speed prediction model in an embodiment of this specification is shown, including:

[0079] S301: Acquire training samples, where the training samples include input samples and sample labels. The input samples include the motor state, initial speed, and throttle command of the sample motor, and the sample labels include the future speed corresponding to the input samples.

[0080] Specifically, for a training sample, the motor state, initial speed, and throttle command of the sample motor can be the control cycle output by the airborne sensor collected under the actual flight state of the target aircraft. The motor status, motor speed, and control cycle of the sample motor The controller outputs the throttle command for the sample motor. The future speed can be the control cycle output by the onboard sensor. The motor speed of the sample motor.

[0081] S303: Input the input sample into the speed prediction model to be trained, and determine a second predicted speed output by the speed prediction model.

[0082] After the input sample is determined, the input sample is input into the speed prediction model to be trained, and the output of the speed prediction model to be trained is used as the sample motor control cycle The second predicted speed.

[0083] In some implementations, the rotational speed prediction model to be trained may be a deep neural network model including at least three hidden layers, thereby ensuring that the rotational speed prediction model has sufficient parameters to extract the nonlinear dynamic characteristics of the target aircraft.

[0084] S305: Training the speed prediction model to be trained according to the difference between the second predicted speed and the ideal speed.

[0085] After the predicted speed is determined, the speed prediction model to be trained can be trained according to the difference between the sample label and the predicted speed and according to a preset loss function.

[0086] Specifically, the preset loss function may be a commonly used loss function such as a cross entropy loss function, a cosine similarity loss, etc., and this specification does not limit this.

[0087] In some implementations, gradient descent or adaptive learning rate algorithms can be used to accelerate the convergence of the loss function during training.

[0088] In some implementations, a simulation dynamics model corresponding to the target aircraft is established based on the control instruction module, motor electronic adjustment model, aerodynamic model, dynamic equations, and kinematic equations corresponding to the target aircraft. The simulation dynamics model is used to simulate the motor speed, motor state, and throttle commands input to each motor of the target aircraft under preset working conditions, and the training samples corresponding to the target aircraft are determined based on the simulation results.

[0089] To ensure the generalization of the speed prediction model, a feasible approach is to train the model using training samples covering a wider and more diverse range of aircraft states. Because the target aircraft's state in real-world flight missions is limited by the actual mission, the limited training samples obtained from real-world missions cannot cover all possible aircraft states.

[0090] Thus, a simulation dynamics model of the target aircraft can be established. The simulation dynamics model is used to simulate the simulated aircraft state of the target aircraft near each preset working condition point, and further determine the motor speed and motor state of each motor in the target aircraft under each simulated working condition and the throttle command input to each motor. For any motor of the target aircraft, the motor is used as a sample motor, and the sample motor is used in any control cycle obtained by simulation, for example, the control cycle Motor speed, motor status and control cycle within Input the throttle command of the sample motor as the input sample, and control the cycle The motor speed of the sample motor is used as the future speed, that is, the sample label, to obtain a training sample.

[0091] The aforementioned operating conditions may include the aircraft state and flight target. The aircraft state may include global angular observations such as the aircraft's position and Earth-axis velocity, as well as more granular observations such as the aircraft's attitude angle and attitude angular velocity. The flight target may include the target position and the velocity at which it will be reached.

[0092] It should be noted that, for special instructions, the target aircraft can operate normally under the aforementioned operating conditions. In other words, the target aircraft has no faults under the signed operating conditions.

[0093] Among them, the simulation dynamics model can be established by using various commonly used simulation model design tools, such as simulink, FLIGHTLAB, JSBSim, etc., and this manual does not impose any restrictions on this.

[0094] In some implementations, when using the simulation dynamics model for simulation, the operating conditions to be simulated can be uniformly selected through an orthogonal experimental design method. This allows a speed prediction model with strong generalization to be trained using fewer training samples.

[0095] In some implementations, the simulation dynamics model can also be used to control the cycle , you can also preset fault conditions to determine the sample motor in the control cycle under fault conditions The fault speed is determined, and the fault threshold corresponding to the motor fault type is determined according to the difference between the fault speed and the future speed.

[0096] Among them, the future speed is the speed of the sample motor in the control cycle under any working condition. The fault condition indicates that the motor speed, motor status, and throttle command input to the sample motor are the same as the operating condition, and the other aircraft states of the target aircraft except for the preset motor fault are the same as the operating condition.

[0097] In some implementations, the speed prediction model is deployed on an onboard computing device of the target aircraft. Thus, the first predicted speed can be determined by using the onboard computing device of the target aircraft without remote communication. On the other hand, the parameter scale of the speed prediction model is much smaller than that of a complete simulation dynamics model. The onboard computing device of the target aircraft can execute a round of predictions in one control cycle. Figure 2 The method shown for determining aircraft power failure has low computational latency.

[0098] It should be understood that the descriptions such as “first” and “second” in this article are only used to distinguish similar concepts for the sake of simplicity of description and do not have any other limiting effect.

[0099] The various embodiments in this specification are described in a progressive manner. Similar parts between the various embodiments can be referred to in conjunction with each other. Each embodiment focuses on the differences between the other embodiments. In particular, the system embodiments are generally similar to the method embodiments, so the description is relatively simple. For relevant parts, refer to the description of the method embodiments.

[0100] The foregoing description of this specification describes specific embodiments. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims can be performed in an order different from that described in the embodiments and still achieve the desired results. Furthermore, the processes depicted in the accompanying drawings do not necessarily require the specific order shown or the sequential order to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0101] Those skilled in the art should further appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the components and steps of each example have been generally described in terms of function in the above description. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art may use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application. Among them, the software module can be placed in a random access memory (RAM), internal memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art.

[0102] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for determining an aircraft power failure, characterized in that: The method is performed using a speed prediction model, which is a machine learning model, and includes: Acquire an initial speed, a motor state, and a throttle command input to a target motor in a target aircraft in a current control cycle as input signals; Determining, using the speed prediction model and based on the input signal, a first predicted speed of the target motor in a next control cycle of the current control cycle; A fault detection result of the target motor is determined according to a difference between the first predicted speed and an actual speed of the target motor after receiving the throttle command.

2. The method according to claim 1, wherein The motor state specifically includes: at least one of motor voltage, motor current and rotor torque load.

3. The method according to claim 1, wherein Determining a fault detection result of the target motor according to a difference between the first predicted speed and an actual speed of the target motor after receiving the throttle command specifically includes: determining a current speed difference according to a difference between the first predicted speed and an actual speed of the target motor after receiving the throttle command; determining a speed loss ratio according to a ratio of the current speed difference to the first predicted speed; A fault detection result of the target motor is determined according to the speed loss ratio and a preset fault threshold.

4. The method according to claim 3, wherein Determining a fault detection result of the target motor according to the speed loss ratio and a preset fault threshold value, specifically comprising: When the speed loss ratio is greater than a preset fault threshold, determining that the target motor is a faulty motor; When the target motor is a faulty motor, determining an expected pulling force of the target motor according to the first predicted speed, and determining an actual pulling force according to the actual speed; According to the difference between the actual pulling force and the expected pulling force, a power loss rate corresponding to the target motor is determined as a fault detection result corresponding to the target motor.

5. The method according to claim 4, wherein After determining the power loss rate corresponding to the target motor, the method further includes: A fault-tolerant control strategy for the target aircraft is determined according to the power loss rate.

6. The method according to claim 4, wherein After determining that the target motor is a faulty motor, the method further includes: A fault detection result of the target motor is determined according to the motor state and the speed loss ratio.

7. The method according to claim 1, wherein Also includes: Acquire a training sample, wherein the training sample includes an input sample and a sample label, wherein the input sample includes a motor state, an initial speed, and a throttle command of a sample motor, and the sample label includes a future speed corresponding to the input sample; Inputting the input sample into the speed prediction model to be trained, and determining a second predicted speed output by the speed prediction model to be trained; The speed prediction model to be trained is trained according to the difference between the second predicted speed and the future speed.

8. The method according to claim 7, wherein Also includes: Establish a simulation dynamics model corresponding to the target aircraft according to the control instruction module, motor and electric control model, aerodynamic model, dynamic equations, and kinematic equations corresponding to the target aircraft; Using the simulation dynamics model, simulating the motor speed, motor state, and throttle command input to each motor of the target aircraft under various preset operating conditions; The training samples corresponding to the target aircraft are determined according to the simulation results.

9. The method according to claim 1, wherein The rotation speed prediction model is deployed on an onboard computing device of the target aircraft.

10. A computing device comprising a memory and a processor, wherein the memory stores executable code, wherein: When the processor executes the executable code, the method according to any one of claims 1 to 9 is implemented.

Images