A method and system for regulating power of a simulation machine actuator cylinder motor, an electronic device and a storage medium

Abstract

The application relates to the computer technical field and is particularly applied to a motion control scene in an artificial intelligence system in a production field. A simulation machine actuating cylinder motor power regulation method and system, an electronic device and a storage medium are disclosed, which comprise the following steps: setting basic parameters and establishing a multi-dimensional sensor error correction model; collecting and calibrating multi-source data such as motors, cockpit attitudes, loads, environments and abnormal sounds in real time; calculating real-time power requirements and power adaptation degrees based on a coupling model of operating actions, loads and power; fusing a tensor self-encoder and a gradient boosting tree to perform power deficiency risk prediction and abnormal sound grading early warning; adopting a multi-modal power regulation mechanism according to the prediction and determination results to implement temperature rise, overload and frequent regulation protection; collecting data after regulation to close loop correct the model and perform incremental updating. The scheme suppresses abnormal sound, improves power matching and system stability, and supports rapid adaptation of different types of simulators.

Description

Technical Field

[0001] This invention relates to the field of computer technology, and specifically to a method, system, electronic device, and storage medium for controlling the power of a simulator actuator motor. Background Technology

[0002] The motion system of a full-motion simulator drives the cockpit to simulate flight attitude via actuators. The power control of the actuator motors directly affects the simulator's simulation experience and operational reliability. In existing technologies, actuator motor power control often employs a fixed power output mode, with some solutions using simple load feedback to achieve coarse adjustment of motor power. For abnormal noise issues in the simulator's motion system, the approach is often reactive, involving post-incident troubleshooting and component replacement. This involves tracing the source through blueprints, measuring circuitry, and identifying cross-connections to locate the faulty component, then replacing it with a more powerful motor or a damaged transmission component. Motor power control methods often rely on manual experience to set or use fixed control parameters, achieving basic control over motor power through a preset parameter system. Some solutions are equipped with basic sensors to collect data related to motor operation, load, and simulator motion, supporting simple power adjustment and troubleshooting, and enabling basic location of faulty actuator motors.

[0003] However, existing technologies have the following drawbacks:

[0004] The power control mode is rigid and not designed for the load differences of multiple operation actions. It cannot match the instantaneous load fluctuations when switching actions, and is prone to power and load mismatch.

[0005] The handling of abnormal noises is a reactive, after-the-fact mode, lacking any advance prediction or active suppression mechanism, and thus failing to address the core cause of power mismatch at its root.

[0006] The control logic lacks adaptive capabilities, relies on human experience or fixed parameters, and cannot cope with dynamic changes in load, environment, and motor performance.

[0007] The sensor uses an error-free design, which makes it susceptible to measurement deviations caused by vibration, temperature, and drift, leading to inaccurate power calculation and control.

[0008] The model has poor adaptability, relying solely on conventional historical data for training, and cannot cope with special working conditions such as extreme angles and complex linkages;

[0009] The fault location capability is weak, and it can only identify faulty motors, but cannot distinguish specific fault types such as motor performance degradation, transmission component wear, and sensor failure; it lacks a full-process motor protection mechanism and has no effective control over overload, overheating, and frequent adjustment.

[0010] It has poor versatility, lacks adaptation solutions for different models of simulators, and is difficult to apply on a large scale; it also lacks a complete closed-loop control system, so the control effect cannot be continuously improved through iteration.

[0011] The system is not adaptable to extreme environments, and its stability and control accuracy decrease under high and low temperatures and strong vibrations.

[0012] To address the aforementioned issues, this invention draws upon the intelligent perception and decision-making approaches in artificial intelligence systems for the production field, combines the modular integration concept of artificial intelligence middleware and function libraries, and utilizes the abnormal noise feature extraction method of computer audiovisual software and the manipulation action recognition technology of biometric recognition software to construct an intelligent power control scheme for simulator actuator motors. This scheme aims to solve the core defects of existing technologies, such as power mismatch, passive handling of abnormal noise, and lack of closed-loop adaptive control. Summary of the Invention

[0013] The purpose of this invention is to provide a method, system, electronic device, and storage medium for controlling the power of the actuator motor in a simulator, in order to solve the technical problems caused by the mismatch between the power and load of the actuator motor in a full-motion simulator under various operating actions, which leads to abnormal noise, as well as the technical problems caused by rigid power control mode, lack of adaptive capability, sensor measurement deviation, poor model adaptability, inaccurate fault location, lack of closed-loop control and universal design.

[0014] To address the aforementioned technical problems, this invention provides a method for controlling the power of a simulator actuator motor, comprising:

[0015] Set basic parameters and establish a multi-dimensional sensor error correction model;

[0016] Multiple types of operating parameters are collected in real time, key data are selected, and real-time error calibration is performed based on the multi-dimensional sensor error correction model to obtain preprocessed real-time operating data.

[0017] Based on the real-time operation data, a coupled mathematical model of multiple control actions-load-power is constructed to calculate the real-time power requirement of the actuator motor, and the power adaptability of the actuator motor is calculated based on the real-time output power of the actuator motor and the real-time power requirement.

[0018] Based on the real-time operating data, the real-time power demand, and the power adaptability, a prediction result of the power insufficiency risk in the future period is obtained through the prediction model, and an abnormal noise warning is output according to the prediction result. At the same time, the power adaptability is monitored to obtain a judgment result on whether the real-time power of the actuator motor is sufficient to support the current operation.

[0019] Based on the prediction results and the judgment results, a multimodal adjustment mechanism is used to adjust the output power of the actuator motor, and the power adaptability and abnormal noise intensity after adjustment are monitored. The parameters of the control model are corrected based on the monitoring results. The control model includes the coupled mathematical model and the prediction model.

[0020] In some specific embodiments, basic parameters are set, and a multi-dimensional sensor error correction model is established, further including:

[0021] The basic parameters include at least:

[0022] Rated power of the actuator motor, angle range of various operating actions, load fluctuation threshold, power deviation threshold, abnormal noise detection threshold, power adaptability threshold, and extreme environment judgment threshold;

[0023] A configurable parameter interface is set up to configure parameters such as actuator stiffness coefficient, transmission mechanism friction coefficient, actuator motor efficiency benchmark value, transmission efficiency benchmark value, and actuator lever arm corresponding to various operating actions according to the simulator model. Parameter templates for multiple simulator models are also built-in.

[0024] By using temperature compensation algorithm, vibration filtering algorithm, and drift calibration algorithm, the sensor error coefficients under different temperature, vibration, and operating length conditions are calibrated, and the multi-dimensional sensor error correction model is established.

[0025] A historical database is built to store historical operating data, including motor power, abnormal noise intensity, power adaptability data, extreme operating condition data, and fault data for various operating actions, different angles, different loads, and different temperatures.

[0026] In some specific embodiments, multiple types of operating parameters are collected in real time, key data are selected, and real-time error calibration is performed based on the multi-dimensional sensor error correction model to obtain preprocessed real-time operating data, which further includes:

[0027] Real-time data acquisition at a set sampling frequency:

[0028] Real-time current, real-time voltage, real-time speed, winding temperature, and running time of the actuator motor; real-time pitch angle, roll angle, yaw angle and corresponding angular velocity and angular acceleration of the simulator cockpit; real-time load of the cockpit, ambient temperature, ambient vibration intensity, and real-time abnormal noise intensity and frequency.

[0029] Based on a dual backup configuration combining a primary sensor and a backup sensor, when the primary sensor fails, it automatically switches to the backup sensor data; when the backup sensor also fails, it supplements the missing data by fitting correlated data.

[0030] The collected data is denoised and abnormal data is removed by time-domain averaging, sliding window integration and median filtering, and real-time error calibration is performed by the multi-dimensional sensor error correction model.

[0031] The calibrated data is standardized to obtain preprocessed real-time operating data.

[0032] In some specific embodiments, based on the real-time operating data, a coupled mathematical model of multiple manipulation actions-load-power is constructed to calculate the real-time power demand of the actuator motor, and the power adaptability of the actuator motor is calculated based on the real-time output power of the actuator motor and the real-time power demand, further including:

[0033] The load composition of various control actions is analyzed. The total load borne by the actuator is calculated by the cockpit gravity load, the inertial load of the control action, and the friction load of the transmission mechanism. Among them, the gravity load and inertial load are calculated differently according to the pitch, roll, yaw and linkage action types, and an environmental vibration correction coefficient is introduced. For linkage actions, the total load is the weighted sum of the loads of each individual action, and an action switching correction coefficient is introduced.

[0034] Based on the total load, the real-time speed of the actuator cylinder, the real-time operating efficiency of the actuator cylinder motor, and the real-time operating efficiency of the transmission mechanism, the real-time power requirement of the actuator cylinder motor is calculated. The real-time speed of the actuator cylinder is obtained by differentiating the displacement of the actuator cylinder with respect to time.

[0035] Based on the real-time operating voltage, real-time operating current and real-time operating efficiency of the actuator motor, the real-time output power of the actuator motor is calculated, and the ratio of the real-time output power to the real-time power demand is used as the power adaptability of the actuator motor. At the same time, the power adaptability change rate is calculated.

[0036] The real-time operating efficiency and the real-time operating efficiency of the transmission mechanism are dynamically corrected based on the operating time of the actuator motor, ambient temperature, winding temperature and ambient vibration intensity.

[0037] The potential correlation features between load and power in multi-source coupled data are extracted by tensor autoencoder, and the model parameters of similar working conditions in historical data are transferred to extreme working conditions by transfer learning algorithm.

[0038] In some specific embodiments, based on the real-time operating data, the real-time power demand, and the power adaptability, a predictive model is used to obtain a prediction result of the power insufficiency risk in future periods, and an abnormal noise warning is output according to the prediction result. Simultaneously, the power adaptability is monitored to determine whether the real-time power of the actuator motor is sufficient to support the current operating action. This further includes:

[0039] The preprocessed real-time running data is constructed into a tensor, and low-dimensional latent space feature vectors are extracted by tensor autoencoder.

[0040] The low-dimensional latent space feature vector is input into the gradient boosting tree model trained by classifying and training the historical database according to the type of manipulation action, the type of environment, and the type of fault, to obtain the predicted value of power deviation, power fit, and fault type label.

[0041] Based on the predicted power deviation, power adaptability, power adaptability change rate, real-time abnormal noise intensity, abnormal noise frequency and environmental parameters, the abnormal noise warning is divided into multiple levels, and based on the correlation between abnormal noise frequency and motor speed, the correlation between abnormal noise frequency and actuator displacement, and sensor working status, preliminary fault type differentiation results are obtained.

[0042] By analyzing historical early warning data, a dynamic correction model for early warning thresholds is established. The thresholds are dynamically adjusted according to environmental parameters and the type of manipulation action. When the power adaptability is lower than the power adaptability threshold, a power insufficiency warning is triggered. When the power adaptability change rate is lower than the set threshold, a power insufficiency aggravation warning is triggered.

[0043] In some specific embodiments, based on the prediction result and the determination result, a multi-modal adjustment mechanism is used to adjust the output power of the actuator motor, further including:

[0044] Based on the real-time power requirements, the base power of the actuator motor is set according to the current operation type and environmental parameters, and power redundancy is reserved in the base power. The base power redundancy value is dynamically adjusted according to the winding temperature, running time and environmental parameters of the actuator motor.

[0045] Based on the warning level, the predicted value of power deviation, the power adaptability, the power adaptability change rate and environmental parameters, the dynamic compensation power and the extreme environment correction power are calculated by the PID control algorithm to obtain the total compensation power adjustment. The control parameters of the PID controller are dynamically adjusted according to the warning level and the environmental type.

[0046] Based on the base power and the total compensation power adjustment, power output is achieved by adjusting the voltage and current of the actuator motor, limiting the power change rate, predicting action switching by analyzing the changing trend of the control action identification data, and adjusting the control parameters in advance.

[0047] Monitor the winding temperature and operating current of the actuator motor. When the winding temperature exceeds the temperature threshold, activate cooling measures or switch to the standby motor. When the real-time output power of the actuator motor exceeds the protection multiple of the rated power of the actuator motor, activate the load shunting mechanism and limit the power regulation frequency of the actuator motor.

[0048] In some specific embodiments, the adjusted power fit and abnormal noise intensity are monitored, and the parameters of the control model are corrected based on the monitoring results, further including:

[0049] Collect the adjusted actuator motor power, abnormal noise intensity, abnormal noise frequency, predicted power deviation, power adaptability, power adaptability change rate, and winding temperature. Combine this with the current operating action type and environmental parameters to determine the adjustment effect and risk status.

[0050] When the adjustment effect is not good or there is a risk, the control parameters of the PID controller are dynamically corrected according to the adjustment effect deviation, the current operation action type, environmental parameters and risk status. The efficiency parameters, load weights of different operation actions and environmental correction coefficients in the coupled mathematical model are corrected, and the risk is handled in a closed loop.

[0051] An incremental training strategy is adopted, which classifies the data according to the type of manipulation action, environment type, and fault type. The adjusted real-time data is added to the historical database to iteratively update the prediction model. When the control performance index deviates from the set threshold, the model is switched to the backup model.

[0052] By combining fault type labels, real-time operating parameters, and abnormal noise characteristics, and comparing historical fault data, the system outputs the location results of the fault location and fault severity, and provides maintenance suggestions.

[0053] Based on the same concept, the present invention also provides a simulator actuator motor power control system, comprising:

[0054] The system initialization and parameter calibration module is configured to set basic parameters and establish a multi-dimensional sensor error correction model.

[0055] The multi-source data real-time acquisition and preprocessing module is configured to acquire multiple types of operating parameters in real time, select key data, perform real-time error calibration based on the multi-dimensional sensor error correction model, and obtain preprocessed real-time operating data.

[0056] The coupling model construction and power demand and adaptability calculation module is configured to construct a coupled mathematical model of multiple operation actions-load-power based on the real-time operation data, calculate the real-time power demand of the actuator motor, and calculate the power adaptability of the actuator motor based on the real-time output power of the actuator motor and the real-time power demand.

[0057] The power insufficiency prediction, abnormal noise warning and power adaptability real-time monitoring module is configured to obtain the prediction result of the power insufficiency risk in the future period through the prediction model based on the real-time operation data, the real-time power demand and the power adaptability, and output the abnormal noise warning according to the prediction result. At the same time, it monitors the power adaptability to obtain the judgment result of whether the real-time power of the actuator motor is sufficient to support the current operation.

[0058] The multimodal power adaptive adjustment and closed-loop feedback module is configured to adjust the output power of the actuator motor using a multimodal adjustment mechanism based on the prediction result and the judgment result, and to monitor the power adaptability and abnormal noise intensity after adjustment, and to correct the parameters of the control model based on the monitoring result, wherein the control model includes the coupled mathematical model and the prediction model.

[0059] Based on the same concept, the present invention also provides an electronic device, including: a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus; the memory stores a computer program, and when the computer program is executed by the processor, the processor performs the steps of a method for power regulation of a simulator actuator motor.

[0060] Based on the same concept, the present invention also provides a computer-readable storage medium storing a computer program executable by an electronic device, which, when run on the electronic device, causes the electronic device to perform the steps of a method for controlling the power of a simulator actuator motor.

[0061] Compared with existing technologies, its advantages are as follows:

[0062] This invention integrates the real-time decision-making and adaptive control concepts of artificial intelligence systems in the production field. It leverages the resource scheduling capabilities of an AI-optimized operating system and the data interaction architecture of AI middleware, combining abnormal noise spectrum analysis from computer audiovisual software with motion pattern recognition methods from biometric software. The invention discloses a method, system, electronic device, and storage medium for controlling the power of a simulator's actuator cylinder motor. Specific beneficial effects are as follows:

[0063] To suppress abnormal noise and ensure power adaptation: By constructing a multi-operation action-load-power coupling mathematical model, the power demand threshold under different operation actions and different operating conditions is quantified, the motor power adaptation degree is calculated in real time, and combined with the early prediction of insufficient power and multi-modal power adaptive adjustment, the core problem of motor power and load mismatch under various operation actions is solved, the generation of abnormal noise is suppressed, and the simulator simulation experience and operation safety are improved.

[0064] Full-process risk control to improve system stability: Real-time compensation for measurement deviations caused by vibration, temperature and drift is achieved through multi-dimensional sensor error correction algorithms. Redundant backup of primary and backup sensors and fitting of related data are used to ensure data continuity. Extreme environment adaptation sub-model is designed to dynamically correct model coefficients. Combined with the full-process motor protection mechanism, effective control of overload, overheating and frequent adjustment is achieved, avoiding various potential risks such as sensor error, data loss, poor stability in extreme environments and increased motor wear.

[0065] Intelligent fault diagnosis reduces maintenance costs: By integrating a composite algorithm of tensor autoencoder and gradient boosting tree, the system predicts the risk of insufficient power and outputs graded abnormal noise warnings. It accurately distinguishes specific fault types such as motor performance degradation, transmission component wear, and sensor failure. Combined with closed-loop feedback control and precise fault location module, it replaces the inefficient troubleshooting methods of traditional post-event inspection and cross-connection of components, improving troubleshooting efficiency and reducing manual inspection and component replacement costs.

[0066] High versatility and adaptability, reducing industrialization costs: Through generalized configuration modules, configurable parameter interfaces and parameter templates for multiple simulator models, there is no need for large-scale modifications to system hardware and algorithms, reducing the cost of solution adaptation.

[0067] Closed-loop optimization and control for continuous iteration: Construct a complete closed-loop control system of "real-time power monitoring - prediction - adjustment - feedback - risk prevention and control - fault handling". Through an incremental model update mechanism, the model parameters are corrected in real time to achieve continuous iterative optimization of the control effect and ensure that the system can maintain the best control performance under different operating conditions.

[0068] Balancing energy consumption optimization and equipment protection: Through variable frequency volumetric speed regulation technology, dynamic correction of basic power, and full-process motor protection, the system reduces energy consumption while ensuring power matching, avoids losses caused by long-term motor overload and frequent adjustments, extends equipment life, and balances control effect with equipment protection. Attached Figure Description

[0069] Other features, objects, and advantages of this application will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings:

[0070] Figure 1 This is a flowchart illustrating some specific embodiments of the simulator actuator motor power control method of the present invention;

[0071] Figure 2 This is a schematic diagram of the structure of a simulator actuator motor power control system in some specific embodiments of the present invention;

[0072] Figure 3 This is a schematic diagram of the structure of an electronic device according to some specific embodiments of the present invention;

[0073] In the diagram, 710 is the processor; 720 is the memory; 730 is the input device; and 740 is the output device. Detailed Implementation

[0074] To make the objectives, technical solutions, and advantages of this application clearer, the application will be further described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0075] The terminology used in the embodiments of this application is for the purpose of describing particular embodiments only and is not intended to limit the application. The singular forms “a,” “said,” and “the” used in the embodiments of this application and the appended claims are also intended to include the plural forms, and “multiple” generally includes at least two unless the context clearly indicates otherwise.

[0076] It should be understood that the term "and / or" used in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Additionally, the character " / " in this article generally indicates that the preceding and following related objects have an "or" relationship.

[0077] It should be understood that although the terms first, second, third, etc., may be used in the embodiments of this application, these descriptions should not be limited to these terms. These terms are only used to distinguish the descriptions. For example, first may also be referred to as second without departing from the scope of the embodiments of this application, and similarly, second may also be referred to as first.

[0078] Depending on the context, the words “if” or “suppose” as used here can be interpreted as “when” or “in response to determination” or “in response to detection.” Similarly, depending on the context, the phrases “if determination” or “if detection (of the stated condition or event)” can be interpreted as “when determination” or “in response to determination” or “when detection (of the stated condition or event)” or “in response to detection (of the stated condition or event).”

[0079] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that an article or device that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such an article or device. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the article or device that includes said element.

[0080] It should be noted that any symbols and / or numbers present in the specification that are not marked in the accompanying drawings are not reference numerals.

[0081] Reference Figure 1 A method for controlling the power of a simulator actuator motor, comprising:

[0082] S101, Set the basic parameters and establish a multi-dimensional sensor error correction model;

[0083] S102, real-time acquisition of multiple types of operating parameters, selection of key data, and real-time error calibration based on the multi-dimensional sensor error correction model to obtain pre-processed real-time operating data;

[0084] S103, Based on the real-time operation data, construct a coupled mathematical model of multiple control actions-load-power, calculate the real-time power requirement of the actuator motor, and calculate the power adaptability of the actuator motor based on the real-time output power of the actuator motor and the real-time power requirement.

[0085] S104, based on the real-time operating data, the real-time power demand, and the power adaptability, a prediction result of the power insufficiency risk in the future period is obtained through the prediction model, and an abnormal noise warning is output according to the prediction result. At the same time, the power adaptability is monitored to obtain a judgment result on whether the real-time power of the actuator motor is sufficient to support the current operation.

[0086] S105, based on the prediction result and the judgment result, a multi-modal adjustment mechanism is used to adjust the output power of the actuator motor, and the power adaptability and abnormal noise intensity after adjustment are monitored. The parameters of the control model are corrected according to the monitoring results, wherein the control model includes the coupled mathematical model and the prediction model.

[0087] Specifically, the system initialization and parameter calibration are performed. Basic parameters are set, including the rated power of the actuator motor, the pitch angle range, roll angle range, yaw angle range, load fluctuation threshold, power deviation threshold, abnormal noise detection threshold, power adaptability threshold, and extreme environment judgment threshold for various control actions. Through temperature compensation algorithms, vibration filtering algorithms, and drift calibration algorithms, sensor error coefficients are calibrated under different temperature, vibration, and operating length conditions, establishing a multi-dimensional sensor error correction model. Simultaneously, a historical database is built to store actuator motor power, abnormal noise intensity, power adaptability data, extreme condition data, and fault data for various control actions, different angles, different loads, and different temperatures. Based on this, a configurable parameter interface is set to configure the actuator stiffness coefficient, transmission mechanism friction coefficient, actuator motor efficiency benchmark value, transmission efficiency benchmark value, and actuator lever arm corresponding to various control actions according to the simulator model, and includes parameter templates for multiple simulator models.

[0088] Real-time acquisition and preprocessing of multi-source data are performed. Multiple operating parameters are acquired in real-time at a set sampling frequency, including real-time current, voltage, speed, winding temperature, and runtime of the actuator motor; real-time pitch, roll, and yaw angles and their corresponding angular velocities and angular accelerations of the simulator cockpit; real-time cockpit load, ambient temperature, and environmental vibration intensity; and real-time abnormal noise intensity and frequency. Time-domain averaging, sliding window integration, and median filtering are used to denoise the acquired data and remove abnormal data caused by sensor failures. A dual-backup design combining a primary and backup sensor is adopted. When the primary sensor fails, data automatically switches to the backup sensor; when the backup sensor also fails, missing data is supplemented by fitting the coupling relationship between angle and displacement. An established multi-dimensional sensor error correction model is used to perform real-time error calibration of key data such as angle, current, and displacement, compensating for measurement deviations caused by temperature, vibration, and drift. The calibrated data is then standardized to obtain preprocessed real-time operating data.

[0089] Based on preprocessed real-time operational data, a multi-control action-load-power coupling mathematical model is constructed. The load composition for various control actions is analyzed, and the total load borne by the actuators is calculated using cockpit gravity load, control action inertial load, and transmission mechanism friction load. Gravity load and inertial load are calculated differently for pitch, roll, yaw, and linked actions, and an environmental vibration correction factor is introduced. For pitch actions, the gravity load is calculated based on the total cockpit mass, gravitational acceleration, and the sine of the pitch angle; the inertial load is calculated based on the total cockpit mass, the actuator lever arm corresponding to the pitch action, and the pitch angular acceleration. Roll and yaw actions are calculated similarly, incorporating the roll and yaw angles and their corresponding lever arms and angular accelerations, respectively. For linked actions, the total load is a weighted sum of the loads of each individual action, and an action switching correction factor is introduced to compensate for instantaneous load fluctuations during action switching. The frictional load is calculated based on the friction coefficient of the transmission mechanism and the support force of the actuator cylinder. The actuator cylinder support force is obtained by coupling the actuator cylinder stiffness coefficient, actuator cylinder displacement, and the cosine value of the corresponding operating angle. Based on the total load, the real-time movement speed of the actuator cylinder, the real-time operating efficiency of the actuator cylinder motor, and the real-time operating efficiency of the transmission mechanism, the real-time power requirement of the actuator cylinder motor is calculated. The real-time movement speed of the actuator cylinder is obtained by differentiating the actuator cylinder displacement with respect to time. Based on the real-time operating voltage, real-time operating current, and real-time operating efficiency of the actuator cylinder motor, the real-time output power of the actuator cylinder motor is calculated. The ratio of this real-time output power to the real-time power requirement is used as the power fit degree of the actuator cylinder motor, and the rate of change of the power fit degree is also calculated. The real-time operating efficiency of the actuator cylinder motor and the real-time operating efficiency of the transmission mechanism are dynamically corrected based on the actuator cylinder motor operating time, ambient temperature, actuator cylinder motor winding temperature, and environmental vibration intensity. The longer the operating time, the higher the ambient temperature, the higher the motor winding temperature, and the greater the vibration intensity, the lower the efficiency. The potential correlation features between load and power in multi-source coupled data are extracted by tensor autoencoder, and the model parameters of similar working conditions in historical data are transferred to extreme working conditions by transfer learning algorithm.

[0090] The system performs power deficiency prediction, abnormal noise warning, and real-time monitoring of power adaptability. Preprocessed real-time operating data is constructed as a tensor. A low-dimensional latent space feature vector is extracted using a tensor autoencoder. This feature vector is input into a gradient boosting tree model trained from a historical database, categorized by manipulation action type, environment type, and fault type. This yields the predicted power deviation, power adaptability, and fault type label. Based on the predicted power deviation, power adaptability, power adaptability change rate, real-time abnormal noise intensity, abnormal noise frequency, and environmental parameters, abnormal noise warnings are classified into multiple levels. When the abnormal noise frequency is positively correlated with motor speed and the motor winding temperature is normal, a preliminary judgment is made that the motor performance has deteriorated; when the abnormal noise frequency is positively correlated with actuator displacement, a preliminary judgment is made that the transmission components are worn; when the sensor's operating status is abnormal and the data fluctuates significantly, a preliminary judgment is made that the sensor is faulty. By analyzing historical early warning data, a dynamic correction model for early warning thresholds is established. The thresholds are dynamically adjusted according to environmental parameters and the type of manipulation action. When the power adaptability is lower than the power adaptability threshold, a power insufficiency warning is triggered. When the power adaptability change rate is lower than the set threshold, a power insufficiency aggravation warning is triggered.

[0091] A multi-modal adjustment mechanism is employed to regulate the output power of the actuator motor. Based on real-time power demand, the base power of the actuator motor is set according to the current operating action type and environmental parameters, with power redundancy reserved within this base power. The base power redundancy value is dynamically adjusted based on the actuator motor winding temperature, operating time, and environmental parameters. Redundancy is reduced when the motor winding temperature exceeds a temperature threshold, increased in extreme environments, and increased when the motor operating time exceeds a duration threshold. Based on the warning level, predicted power deviation, power adaptability, power adaptability change rate, and environmental parameters, a PID control algorithm calculates the dynamic compensation power and extreme environment correction power to obtain the total compensation power adjustment. The proportional, integral, and derivative coefficients of the PID controller are dynamically adjusted according to the warning level and environmental type. Based on the base power and total compensation power adjustment, power output is achieved by adjusting the voltage and current of the actuator motor, while limiting the power change rate to introduce an instantaneous power buffer mechanism. By analyzing the angular velocity and angular acceleration change rate of the operating action recognition data, the action switching trend is predicted, and the control parameters of the PID controller are adjusted in advance. The system monitors the temperature of the actuator motor windings and the motor operating current. When the motor winding temperature exceeds the first temperature threshold, cooling measures are initiated. When it exceeds the second temperature threshold, the system switches to the standby motor. When the real-time output power of the motor exceeds the protection multiple of the motor's rated power, the load shunting mechanism is initiated, and the frequency of the actuator motor power regulation is limited.

[0092] The system implements closed-loop feedback control and model parameter correction. It collects data on the adjusted actuator motor power, abnormal noise intensity, abnormal noise frequency, predicted power deviation, power adaptability, rate of change of power adaptability, and actuator motor winding temperature. Combined with the current control action type and environmental parameters, it assesses the control effect and risk status. When the control effect is poor or risks exist, it dynamically corrects the proportional, integral, and derivative coefficients of the PID controller based on the control effect deviation, current control action type, environmental parameters, and risk status. It also corrects the efficiency parameters in the coupled mathematical model, the load weights for different control actions, and the environmental correction coefficients. Closed-loop handling is implemented for risks such as motor overheating, sensor malfunction, extreme environment, motor performance degradation, and transmission component wear. An incremental training strategy is adopted, categorized by control action type, environmental type, and fault type. The adjusted real-time data is added to the historical database, iteratively updating the prediction model. When control performance indicators deviate from the set threshold, it switches to a preset backup model. Combining fault type labels, real-time operating parameters, and abnormal noise characteristics with historical fault data, it outputs the location and severity of the fault and provides maintenance suggestions.

[0093] In some applications, basic parameters are set, and a multi-dimensional sensor error correction model is established. The basic parameters include at least the rated power of the actuator motor, the angle range of various operating actions, the load fluctuation threshold, the power deviation threshold, the abnormal noise detection threshold, the power adaptability threshold, and the extreme environment judgment threshold. A configurable parameter interface is set to configure parameters such as the actuator stiffness coefficient, the transmission mechanism friction coefficient, the actuator motor efficiency benchmark value, the transmission efficiency benchmark value, and the actuator lever arm corresponding to various operating actions according to the simulator model, and multiple simulator parameter templates are built in. The sensor error coefficients under different temperature, vibration, and running length conditions are calibrated through temperature compensation algorithm, vibration filtering algorithm, and drift calibration algorithm to establish the multi-dimensional sensor error correction model. A historical database is built to store historical operating data including actuator motor power, abnormal noise intensity, power adaptability data, extreme operating condition data, and fault data for various operating actions, different angles, different loads, and different temperatures.

[0094] Understandably, basic parameters are set, including the rated power of the actuator motor, the angle range of various control actions, load fluctuation threshold, power deviation threshold, abnormal noise detection threshold, power adaptability threshold, and extreme environment judgment threshold. The angle range of various control actions corresponds to the upper and lower limits of pitch, roll, and yaw angles, respectively. The load fluctuation threshold is used to determine if there are abnormal load fluctuations. The power deviation threshold is used to determine if the deviation between the actual power and the required power exceeds the allowable range. The abnormal noise detection threshold is used to determine if abnormal noise is generated. The power adaptability threshold is used to determine if the motor's real-time power is sufficient to support the current control action. The extreme environment judgment threshold is used to identify extreme operating conditions such as high and low temperatures and strong vibrations. A configurable parameter interface is set to configure the actuator stiffness coefficient, transmission mechanism friction coefficient, actuator motor efficiency benchmark value, transmission efficiency benchmark value, and actuator lever arm corresponding to various control actions according to the simulator model. It also includes parameter templates for multiple simulator models, supporting one-click import and export of parameters, allowing for rapid adaptation to different simulator models without large-scale modifications to the system hardware and core algorithms. By employing temperature compensation, vibration filtering, and drift calibration algorithms, sensor error coefficients are calibrated under different temperatures, vibration frequencies, and operating durations, establishing a multi-dimensional sensor error correction model. This model can compensate in real time for the impact of temperature changes, vibration interference, and long-term operational drift on sensor measurement accuracy, ensuring the accuracy of subsequent data acquisition. A historical database is constructed to store historical operating data, including actuator motor power, abnormal noise intensity, power adaptability data, extreme operating condition data, and fault data under various operating actions, angles, loads, and temperatures. This provides data support for subsequent model training, parameter optimization, and fault diagnosis.

[0095] In some applications, multiple operating parameters are collected in real time, key data are selected, and real-time error calibration is performed based on the multi-dimensional sensor error correction model to obtain preprocessed real-time operating data. This includes real-time acquisition at a set sampling frequency of: real-time current, real-time voltage, real-time speed, winding temperature, and running time of the actuator motor; real-time pitch angle, roll angle, yaw angle and corresponding angular velocity and angular acceleration of the simulator cockpit; real-time cockpit load, ambient temperature, and ambient vibration intensity; and real-time abnormal noise intensity and frequency. Based on a dual-backup configuration combining primary and backup sensors, when the primary sensor fails, the system automatically switches to backup sensor data; when the backup sensor also malfunctions, missing data is supplemented through correlation data fitting. The collected data is denoised and abnormal data is removed based on time-domain averaging, sliding window integration, and median filtering. Real-time error calibration is performed using the multi-dimensional sensor error correction model. The calibrated data is then standardized to obtain preprocessed real-time operating data.

[0096] Understandably, multiple operating parameters are collected in real time at a set sampling frequency, including real-time current, real-time voltage, real-time speed, winding temperature, and running time of the actuator motor; real-time pitch angle, roll angle, yaw angle, and their corresponding angular velocity and angular acceleration of the simulator cockpit; real-time cockpit load, ambient temperature, and ambient vibration intensity; and real-time abnormal noise intensity and frequency. Among these, current, voltage, and speed are used to calculate the motor's real-time output power; winding temperature and running time are used to correct motor operating efficiency; angles and their angular velocities and angular accelerations are used to identify the type of control action and calculate inertial load; cockpit load is used to calculate gravity load; ambient temperature and vibration intensity are used to correct sensor measurement deviations and model parameters; and abnormal noise intensity and frequency are used for abnormal noise warnings and fault type differentiation. To ensure the continuity and reliability of data acquisition, a dual-backup configuration combining primary and backup sensors is adopted. When a primary sensor fails, the control system automatically switches to the corresponding backup sensor data stream, ensuring uninterrupted data acquisition. When the backup sensor also malfunctions, the system supplements missing data through correlation data fitting, for example, by utilizing the kinematic coupling relationship between angle and displacement and extrapolating the missing data values ​​based on data from other normally functioning sensors. The acquired raw data undergoes denoising processing using a combination of time-domain averaging, sliding window integration, and median filtering to eliminate abnormal data points caused by momentary sensor failures or external interference, thus smoothing the data waveform. The denoised data is input into an established multi-dimensional sensor error correction model. Based on the currently acquired ambient temperature, vibration intensity, and motor operating time, the model dynamically calculates the error coefficients of each sensor under the given operating conditions, performing real-time error calibration on key data such as angle, current, and displacement to compensate for measurement deviations caused by temperature changes, vibration interference, and long-term operational drift. The calibrated data is standardized to map data of different dimensions to a preset numerical range, eliminating the impact of dimensional differences on subsequent model calculations and obtaining pre-processed real-time operating data, which provides an accurate data foundation for subsequent coupled model construction, power demand calculation and predictive analysis.

[0097] In some applications, based on the real-time operating data, a coupled mathematical model of multiple control actions, loads, and power is constructed to calculate the real-time power requirements of the actuator motor. The power adaptability of the actuator motor is then calculated based on its real-time output power and the real-time power requirements. This includes analyzing the load composition during various control actions, calculating the total load borne by the actuator through cockpit gravity load, control action inertial load, and transmission mechanism friction load. Gravity load and inertial load are calculated differently based on pitch, roll, yaw, and linked action types, and an environmental vibration correction coefficient is introduced. For linked actions, the total load is a weighted sum of the loads of each individual action, and an action switching correction coefficient is introduced. Based on the total load, the real-time speed of the actuator, the real-time operating efficiency of the actuator motor, and the transmission... The mechanism's real-time operating efficiency is calculated by determining the real-time power requirement of the actuator motor, where the real-time speed of the actuator is obtained by differentiating the actuator displacement with respect to time. Based on the actuator motor's real-time operating voltage, real-time operating current, and real-time operating efficiency, the actuator motor's real-time output power is calculated, and the ratio of the real-time output power to the real-time power requirement is used as the actuator motor's power fit. The rate of change of the power fit is also calculated. The real-time operating efficiency and the transmission mechanism's real-time operating efficiency are dynamically corrected based on the actuator motor's operating time, ambient temperature, winding temperature, and ambient vibration intensity. Potential correlation features between load and power in multi-source coupled data are extracted using a tensor autoencoder, and a transfer learning algorithm is used to transfer model parameters from similar operating conditions in historical data to extreme operating conditions.

[0098] Understandably, the load composition during various control actions is analyzed, and the total load borne by the actuator is calculated using the cockpit gravity load, the inertial load of the control action, and the frictional load of the transmission mechanism. The gravity load and inertial load are calculated differently based on the type of pitch, roll, yaw, and linked actions. For pitch actions, the gravity load is calculated based on the total cockpit mass, gravitational acceleration, and the sine of the pitch angle; the inertial load is calculated based on the total cockpit mass, the actuator lever arm corresponding to the pitch action, and the pitch angle acceleration. For roll actions, the gravity load is calculated based on the total cockpit mass, gravitational acceleration, and the sine of the roll angle; the inertial load is calculated based on the total cockpit mass, the actuator lever arm corresponding to the roll action, and the roll angle acceleration. For yaw actions, the gravity load is calculated based on the total cockpit mass, gravitational acceleration, and the sine of the yaw angle; the inertial load is calculated based on the total cockpit mass, the actuator lever arm corresponding to the yaw action, and the yaw angle acceleration. The above calculations incorporate an environmental vibration correction factor to compensate for the impact of vibration on load measurement and calculation. For linked actions, the total load is the weighted sum of the loads of each individual action. The weight of each individual action is determined based on its amplitude proportion in the linked action. An action switching correction factor is also introduced to compensate for dynamic load abrupt changes during multi-action switching. The frictional load is calculated based on the transmission mechanism's friction coefficient and the actuator cylinder's support force. The actuator cylinder's support force is obtained by coupling the actuator cylinder stiffness coefficient, actuator cylinder displacement, and the cosine value of the corresponding operating angle. Based on the total load, the actuator cylinder's real-time speed, the actuator cylinder motor's real-time operating efficiency, and the transmission mechanism's real-time operating efficiency, the actuator cylinder motor's real-time power requirement is calculated. The actuator cylinder's real-time speed is obtained by differentiating the actuator cylinder displacement with respect to time. Based on the actuator cylinder motor's real-time operating voltage, real-time operating current, and real-time operating efficiency, the actuator cylinder motor's real-time output power is calculated. The ratio of this real-time output power to the real-time power requirement is used as the actuator cylinder motor's power fit. Simultaneously, the power fit rate of change is calculated to characterize the dynamic trend of the power fit. The real-time operating efficiency of the actuator motor and the transmission mechanism is dynamically adjusted based on the actuator motor's operating time, ambient temperature, winding temperature, and environmental vibration intensity. The longer the motor's operating time, the higher the ambient temperature, the higher the motor winding temperature, and the greater the environmental vibration intensity, the lower the operating efficiency. High-dimensional feature compression is performed on multi-source coupled data using a tensor autoencoder to extract potential correlation features between load and power under different operating actions, environments, and fault types. A transfer learning algorithm is then used to transfer model parameters from similar operating conditions in historical data to extreme operating conditions, ensuring the accuracy of power demand calculations under special conditions such as extreme angles and complex linkages.

[0099] In some applications, based on the real-time operating data, the real-time power demand, and the power adaptability, a predictive model is used to obtain a prediction result of the power insufficiency risk in the future period. An abnormal noise warning is output based on the prediction result, and the power adaptability is monitored to determine whether the real-time power of the actuator motor is sufficient to support the current operation. This includes constructing a tensor from the preprocessed real-time operating data, extracting low-dimensional latent space feature vectors using a tensor autoencoder, and inputting the low-dimensional latent space feature vectors into a gradient boosting tree model trained by classifying historical databases according to operation type, environment type, and fault type to obtain a power deviation prediction. The system includes a judgment value, power adaptability, and fault type label. Based on the predicted power deviation, power adaptability, power adaptability change rate, real-time abnormal noise intensity, abnormal noise frequency, and environmental parameters, abnormal noise warnings are divided into multiple levels. Preliminary fault type differentiation results are obtained based on the correlation between abnormal noise frequency and motor speed, the correlation between abnormal noise frequency and actuator displacement, and sensor operating status. A dynamic correction model for warning thresholds is established by analyzing historical warning data. The thresholds are dynamically adjusted according to environmental parameters and operation action type. When the power adaptability is lower than the power adaptability threshold, a power insufficiency warning is triggered. When the power adaptability change rate is lower than the set threshold, a power insufficiency aggravation warning is triggered.

[0100] Understandably, the preprocessed real-time operational data is constructed as a tensor. This tensor is organized dimensionally as follows: the first dimension is the variable dimension, encompassing core variables such as motor current, voltage, speed, winding temperature, operating time, pitch angle, roll angle, yaw angle, angular velocity, angular acceleration, cabin load, ambient temperature, ambient vibration intensity, abnormal noise intensity, and abnormal noise frequency; the second dimension is the spatial location dimension, corresponding to the installation positions of each sensor on the simulator; and the third dimension is the time step dimension, containing data from multiple consecutive sampling periods. This tensor is encoded using a tensor autoencoder to extract low-dimensional latent space feature vectors. These feature vectors capture the spatiotemporal coupling characteristics under multiple variables, multiple spatial points, and multiple time steps, while avoiding information redundancy and loss of key features. The extracted low-dimensional latent space feature vectors are input into a prediction model, which is constructed using a gradient boosting tree algorithm and categorized according to control action type, environment type, and fault type. The model is trained using labeled data from a historical database. The model outputs three results: a predicted power deviation, i.e., the predicted deviation between the actual output power of the motor and the required power in the future period; the current calculated value of the power fit; and a fault type label, used to identify possible fault types such as motor performance degradation, transmission component wear, and sensor failure. Based on the model's output of the predicted power deviation, power fit, power fit rate of change, real-time collected abnormal noise intensity and frequency, and current environmental parameters, abnormal noise warnings are divided into multiple levels, each corresponding to different degrees of power insufficiency and control response intensity. Simultaneously, based on the correlation between abnormal noise frequency and motor speed, the correlation between abnormal noise frequency and actuator displacement, and the sensor's operating status, the fault type is preliminarily distinguished: if the abnormal noise frequency shows a positive correlation with motor speed and the motor winding temperature is normal, it is preliminarily determined to be motor performance degradation; if the abnormal noise frequency shows a positive correlation with actuator displacement, it is preliminarily determined to be transmission component wear; if the sensor's operating status shows abnormalities and the data fluctuates significantly, it is preliminarily determined to be sensor failure. A dynamic correction model for warning thresholds is established by analyzing historical warning data. This model dynamically adjusts the power adaptability threshold and the power adaptability change rate threshold based on current environmental parameters and the type of operation, avoiding false warnings caused by data fluctuations or environmental interference. When the real-time monitored power adaptability falls below the dynamically adjusted power adaptability threshold, a power insufficiency warning is triggered; when the power adaptability change rate continues to decline and falls below the dynamically adjusted change rate threshold, a power insufficiency aggravation warning is triggered, indicating that the power insufficiency is worsening.

[0101] In some applications, based on the predicted and determined results, a multimodal adjustment mechanism is used to adjust the output power of the actuator motor. This includes setting the base power of the actuator motor based on the real-time power demand, combined with the current operating action type and environmental parameters, reserving power redundancy in the base power, and dynamically correcting the base power redundancy value according to the actuator motor winding temperature, running time, and environmental parameters. Based on the warning level, predicted power deviation, power adaptability, power adaptability change rate, and environmental parameters, a PID control algorithm is used to calculate the dynamic compensation power and extreme environment correction power to obtain the total compensation power adjustment. In this system, the control parameters of the PID controller are dynamically adjusted according to the warning level and environmental type. Based on the base power and the total compensation power adjustment, the power output is achieved by adjusting the voltage and current of the actuator motor, limiting the power change rate. The action switching is predicted by analyzing the change trend of the control action identification data, and the control parameters are adjusted in advance. The winding temperature and operating current of the actuator motor are monitored. When the winding temperature exceeds the temperature threshold, cooling measures are initiated or the motor is switched to a standby motor. When the real-time output power of the actuator motor exceeds the protection multiple of the rated power of the actuator motor, the load shunting mechanism is initiated, and the power adjustment frequency of the actuator motor is limited.

[0102] Understandably, based on real-time power requirements, the base power of the actuator motor is set according to the current operating action type and environmental parameters, with a power redundancy reserved in the base power to cope with instantaneous load fluctuations during action switching. The base power redundancy value is not fixed but dynamically adjusted based on the actuator motor winding temperature, operating time, and environmental parameters: when the motor winding temperature is too high, the redundancy value is appropriately reduced to avoid overheating; when in extreme high or low temperature or strong vibration environments, the redundancy value is appropriately increased to enhance system stability; when the motor operating time exceeds the set value, the redundancy value is appropriately increased to compensate for motor performance degradation. Based on the warning level, predicted power deviation, power adaptability, power adaptability change rate, and environmental parameters, the dynamic compensation power and extreme environment correction power are calculated using a PID control algorithm to obtain the total compensation power adjustment. The proportional, integral, and derivative coefficients of the PID controller are not fixed values ​​but dynamically adjusted according to the warning level and environmental type: the higher the warning level, the larger the control parameter values, and the stronger the adjustment response; in extreme environments, the control parameters are further increased based on the corresponding warning level to ensure adjustment accuracy. Dynamic compensation power is primarily used to compensate for power deviations and insufficient adaptability, while extreme environment correction power is used to compensate for the impact of environmental parameters deviating from standard operating conditions on system performance. Based on the adjustment of the base power and total compensation power, power output is achieved by adjusting the voltage and current of the actuator motor, while limiting the rate of power change and introducing an instantaneous power buffer mechanism to avoid sudden power surges affecting the transmission mechanism and motor. By analyzing the changing trends of angular velocity and angular acceleration in the manipulation action recognition data, the occurrence of action switching can be predicted in advance, and the control parameters of the PID controller can be adjusted in advance before the action switching to optimize the adjustment response speed and avoid instantaneous power shortages caused by control lag. The system monitors the winding temperature and operating current of the actuator motor in real time. When the winding temperature exceeds the first temperature threshold, cooling measures such as air cooling or liquid cooling are activated. When the winding temperature exceeds the second temperature threshold, the system automatically switches to the backup motor to prevent the main motor from overheating and being damaged. When the real-time output power of the motor exceeds the protection multiple of the motor's rated power, the system activates the load shunting mechanism to coordinate other actuator motors under the same operating action to share the load and prevent a single motor from being overloaded for a long time. At the same time, the system limits the power adjustment frequency of the actuator motor to avoid frequent adjustments that aggravate motor wear. When the motor's running time exceeds the set value, the adjustment frequency is further reduced to extend the motor's service life.

[0103] In some applications, the adjusted power fit and abnormal noise intensity are monitored. Based on the monitoring results, the parameters of the control model are corrected. This includes collecting the adjusted actuator motor power, abnormal noise intensity, abnormal noise frequency, predicted power deviation, power fit, power fit change rate, and winding temperature. Combined with the current operation type and environmental parameters, the adjustment effect and risk status are judged. When the adjustment effect is poor or there is a risk, the control parameters of the PID controller are dynamically corrected based on the adjustment effect deviation, the current operation type, environmental parameters, and risk status. The efficiency parameters, load weights of different operation actions, and environmental correction coefficients in the coupled mathematical model are corrected, and the risk is handled in a closed loop. An incremental training strategy is adopted, which classifies the data by operation type, environment type, and fault type. The adjusted real-time data is added to the historical database to iteratively update the prediction model. When the control performance index deviates from the set threshold, the model is switched to the backup model. Combining the fault type label, real-time operating parameters, and abnormal noise characteristics, the model is compared with historical fault data to output the location results of the fault location and fault degree, and output maintenance suggestions.

[0104] Understandably, the system collects and adjusts the actuator motor power, abnormal noise intensity, abnormal noise frequency, predicted power deviation, power adaptability, power adaptability change rate, and winding temperature. This data, combined with the current operating action type and environmental parameters, is used to comprehensively assess the adjustment effect and risk status. The criteria for judging the adjustment effect include: whether the adjusted power deviation exceeds the power deviation threshold, whether the power adaptability is lower than the power adaptability threshold, and whether the abnormal noise intensity exceeds the abnormal noise detection threshold. The criteria for judging the risk status include: whether the motor winding temperature exceeds the temperature threshold, whether the motor efficiency is continuously lower than the efficiency threshold, whether the friction of the transmission components is continuously increasing, and whether the sensor operating status is abnormal. When the judgment result indicates poor adjustment effect or risk, the relevant parameters in the control model are dynamically adjusted based on the adjustment effect deviation, the current operating action type, environmental parameters, and risk status. Specifically, this includes: correcting the proportional, integral, and derivative coefficients of the PID controller to better adapt the adjustment response to the current operating conditions; correcting motor efficiency parameters, transmission efficiency parameters, load weights for different operating actions, and environmental correction coefficients in the coupled mathematical model to ensure the accuracy of power demand calculations; and initiating corresponding closed-loop handling measures for risks such as motor overheating, sensor malfunction, extreme environment, motor performance degradation, and transmission component wear, such as increasing cooling efforts, switching to backup sensors, adjusting extreme environment correction coefficients, issuing maintenance warnings, and reducing adjustment frequency. An incremental training strategy is adopted, categorizing data by operating action type, environmental type, and fault type, adding the adjusted real-time data to the historical database, and iteratively updating the predictive model to continuously adapt to new operating conditions. Simultaneously, when control performance indicators deviate from the set threshold, the system automatically switches to a preset backup model to ensure the stability of prediction and adjustment. Finally, by combining the fault type labels, real-time operating parameters, and abnormal noise characteristics output by the model, and comparing them with the fault data in the historical database, the fault location and severity are accurately determined. The model outputs a location result that includes the specific faulty motor number, faulty sensor number, faulty transmission component, and targeted maintenance suggestions, providing clear maintenance guidance for operators.

[0105] The following describes another embodiment of the power control method for a simulator actuator motor according to the present invention:

[0106] This embodiment includes:

[0107] System initialization and parameter calibration:

[0108] Initialize the actuator motors, sensors, and controllers of the full-motion simulator's motion system, clarifying the installation position of each actuator and its corresponding control action (highlighting the actuators corresponding to pitch, roll, yaw, and linked actions). Calibrate sensor accuracy (including current sensors, voltage sensors, speed sensors, angle sensors (covering pitch, roll, and yaw angles), actuator displacement sensors, temperature sensors, control action recognition sensors, and sound sensors). Add a new full-dimensional sensor error calibration step, using temperature compensation algorithms (covering extreme temperature ranges from -40℃ to 85℃), vibration filtering algorithms (suppressing vibration interference from 10 to 1000Hz), and drift calibration algorithms (real-time correction of long-term sensor drift). Pre-calibrate sensor error coefficients under different temperature, vibration, and operating length conditions, and establish a multi-dimensional error correction model for subsequent real-time data calibration. Simultaneously, perform periodic automatic calibration (every 24 hours) to mitigate sensor measurement error risks.

[0109] Set the system's basic parameters, including: motor rated power. Angle range for various control actions: pitch ,in This is the minimum angle threshold for pitch motion. The maximum angle threshold for pitch motion, roll ,in Set a minimum angle threshold for horizontal scrolling. Set the maximum angle threshold for roll and yaw. in This represents the minimum angle threshold for yaw action. Maximum yaw angle threshold, angular velocity threshold, and load fluctuation threshold. (Load fluctuation threshold that triggers load anomaly detection), power deviation threshold (When the deviation between actual power and required power exceeds this threshold, it is judged as a risk of insufficient power), abnormal noise detection threshold. (Abnormal noise intensity is detected by a sound sensor; noise exceeding this threshold is considered abnormal noise.) Power compatibility threshold. ( , A power level of ≥0.9 is considered sufficient to support the current control action. A value <0.9 indicates insufficient power adaptation, triggering a warning; extreme environment thresholds (high and low temperatures, strong vibration criteria); a universal parameter configuration interface, which can be configured according to the actuator stiffness coefficient of different simulator models. (Rigidity characteristic parameters of actuator cylinder structure), coefficient of friction (Motion friction characteristic parameters of transmission mechanism), motor efficiency benchmark value (Rated motor efficiency under standard operating conditions), transmission efficiency benchmark value (Rated efficiency of the transmission mechanism under standard operating conditions), the actuator arm corresponding to pitch / roll / yaw motion. Parameters such as the vertical rotation radius of the shaft corresponding to the force application point of the actuator and the structural calibration parameters of the simulator can be flexibly configured. The simulator has built-in parameter templates for multiple models, reducing adaptation costs and avoiding the risk of insufficient universality. Extreme environment parameter calibration allows for the pre-calibration of model correction coefficients under different extreme environments, laying the foundation for extreme environment adaptation.

[0110] Simultaneously, historical operating data of the simulator (including motor power, abnormal noise intensity, and power adaptability data under different control actions, angles, loads, and temperatures) is collected, supplemented with extreme operating condition data (such as extreme pitch / roll / yaw angles, complex linkage actions, high and low temperature environments, and strong vibration environments) and fault data (operating data corresponding to motor performance degradation, transmission component wear, and sensor failures) to build a comprehensive historical database for model training and parameter optimization. By referring to historical fault data of similar simulators, the adaptability of the model and the accuracy of fault identification are improved.

[0111] Real-time acquisition of multi-source data for multiple manipulation actions:

[0112] When the pilot performs various control maneuvers (pitch, roll, yaw, and combined maneuvers), multi-dimensional operational data is collected in real time through various sensors at a sampling frequency of 100Hz. The collected data includes:

[0113] Motor operating parameters: Real-time current of each actuator motor Real-time voltage Real-time rotation speed Motor winding temperature Motor running time ;

[0114] Control actions and motion parameters: Real-time pitch angle of the simulator cockpit Roll angle Yaw angle Corresponding angular velocity angular acceleration In addition, data on the type of manipulation action (obtained through manipulation action recognition sensors to determine whether the current action is a single action or a linked action, and the trend of action switching).

[0115] Load and environmental parameters: Real-time cockpit load Ambient temperature Environmental vibration intensity Real-time displacement of the actuator cylinder ;

[0116] Abnormal noise detection parameters: Real-time abnormal noise intensity collected by a sound sensor Abnormal noise frequency ;

[0117] Sensor operating parameters: the working status of each sensor.

[0118] The collected data undergoes preprocessing: time-domain averaging, sliding window integration, and median filtering are used for noise reduction to remove abnormal data (such as abrupt changes caused by sensor malfunctions); a multi-dimensional sensor error correction model is used to perform real-time error calibration on key data such as angle, current, and displacement to compensate for measurement deviations caused by temperature, vibration, and drift; a new abnormal redundancy backup mechanism is added, adopting a dual backup design of "main sensor + backup sensor". When a main sensor fails, the system automatically switches to backup sensor data. If the backup sensor also fails, it is supplemented by fitting other related data (such as the coupling relationship between angle and displacement) to avoid control failures caused by data loss; extreme environment data preprocessing is performed to specifically correct data collected under extreme high and low temperatures and strong vibration environments to improve data reliability; finally, standardization is performed to map the data to the [0,1] interval to avoid the influence of dimensions. Preliminary data processing is completed through the edge computing module to improve data transmission and processing efficiency, while the preprocessed data is backed up locally to prevent data loss.

[0119] Construction of a multi-manipulation action-load-power coupling model and calculation of power demand and fit:

[0120] Based on the collected preprocessed (including calibration) data, a multi-operation action-load-power coupling mathematical model is constructed to accurately calculate the real-time power demand of the actuator motor under different operation actions, working conditions, and environments. Simultaneously calculate the motor power matching degree in real time. It monitors whether the real-time power of the motor is sufficient to support the current operation, associates fault type characteristics, and provides a basis for fault location.

[0121] First, let's analyze the load composition during various control actions: the load on the actuator mainly includes the cockpit gravity load. Inertial load of manipulation action Frictional load of transmission mechanism The formula for calculating the total load is:

[0122] ;

[0123] in,

[0124] : The total load borne by the actuator at all times;

[0125] Real-time variables;

[0126] : Constant cabin gravity load;

[0127] : Constantly manipulate the inertial load of the action;

[0128] : The transmission mechanism is constantly subjected to frictional load.

[0129] The calculation methods for gravity load and inertial load are designed differently based on the type of manipulation action, and an environmental vibration correction factor is also introduced. ( The greater the vibration intensity, (The larger the value), the better the compensation for the impact of vibration on the load:

[0130] Pitching motion:

[0131] Cockpit gravity load: ;

[0132] Inertial load: ;

[0133] Horizontal rolling action:

[0134] Cockpit gravity load: ;

[0135] Inertial load: ;

[0136] Yaw maneuvers:

[0137] Cockpit gravity load: ;

[0138] Inertial load: ;

[0139] in, Environmental vibration correction factor;

[0140] The total mass of the simulator cockpit and its payload;

[0141] : Gravitational acceleration;

[0142] : respectively Real-time pitch, roll, and yaw angles of the cockpit;

[0143] : respectively The cockpit's pitch, roll, and yaw acceleration at all times.

[0144] These are the actuator arms corresponding to pitch, roll, and yaw movements, respectively.

[0145] Linked actions: The total load is the weighted sum of the loads of the corresponding individual actions (weights) : No. The percentage of amplitude of a single action within a linked action. , (This refers to the number of individual actions included in a linked action), and an action switching correction coefficient is introduced. ( The faster the action switching speed, The larger the value, the better (used to compensate for dynamic load fluctuations during multi-action switching), and to compensate for instantaneous load fluctuations during action switching.

[0146] Frictional load:

[0147] in, The friction coefficient of the transmission mechanism can be determined based on the motor's operating time. Dynamic correction The longer, The larger the value, the better it is used to judge the degree of wear of transmission components;

[0148] The constant force of the actuator cylinder is due to the displacement of the actuator cylinder. It is obtained by coupling calculation with the corresponding manipulation angle.

[0149] (Looking up and down);

[0150] (roll);

[0151] (yaw);

[0152] This is the stiffness coefficient of the actuator cylinder.

[0153] Secondly, a coupled model of load and power is constructed, taking into account motor efficiency. Transmission efficiency The formula for calculating real-time power demand is:

[0154] ;

[0155] in,

[0156] : Real-time power demand of the actuator motor;

[0157] : The real-time speed of the actuator is determined by the displacement of the actuator. Taking the derivative with respect to time, we get

[0158] ;

[0159] Real-time operating efficiency of the motor;

[0160] Real-time operating efficiency of the transmission mechanism;

[0161] Displacement differential;

[0162] Time derivative;

[0163] At the same time, the motor power matching degree is calculated in real time. The formula is:

[0164] ;

[0165] in, : Motor power compatibility at all times;

[0166] : The real-time output power of the motor is determined by... Calculated;

[0167] : Real-time operating voltage of the motor;

[0168] for Real-time operating current of the motor;

[0169] When the value is 0.9, it is determined that the real-time power of the motor is sufficient to support the current operating action; When this occurs, it is determined that the power adaptation is insufficient, triggering a power insufficiency warning; at the same time, the power adaptation change rate is calculated. This is used to characterize the dynamic trend of power fit. The power shortage continued to decline, and the trend of insufficient power was anticipated to worsen in advance.

[0170] Based on historical databases, a tensor autoencoder is used to perform high-dimensional feature compression on multi-source coupled data, extracting potential correlation features between load and power under different operating actions, environments, and fault types, and optimizing the coupling model parameters. The extreme condition adaptation module uses a transfer learning algorithm to transfer model parameters from similar conditions in historical data to extreme conditions, quickly completing model adaptation. At the same time, combined with extreme environment correction coefficients, the accuracy of power demand and power fit calculations under extreme conditions and environments is improved. A new fault feature association module is added to associate power fit, abnormal noise frequency, motor operating parameters, and fault types (motor performance degradation, transmission component wear, sensor failure), providing a basis for subsequent fault location.

[0171] motor efficiency Based on motor running time Ambient temperature Environmental vibration intensity and motor winding temperature Dynamic correction is performed, and the correction formula is as follows:

[0172] ;

[0173] in, Real-time motor efficiency after multi-factor correction;

[0174] Real-time ambient temperature;

[0175] Real-time temperature of motor windings;

[0176] : Cumulative running time of the motor;

[0177] Real-time environmental vibration intensity;

[0178] Under standard operating conditions (ambient temperature 25℃, motor winding temperature 75℃, cumulative running time 0h, vibration intensity...), The rated efficiency benchmark value of the motor;

[0179] : The benchmark value of environmental vibration intensity under standard operating conditions.

[0180] The correction formula is similar to the motor efficiency correction formula, but with the addition of corrections for running time and vibration intensity, the formula is:

[0181] ;

[0182] in, The rated efficiency benchmark value of the transmission mechanism under standard operating conditions.

[0183] Insufficient power prediction, abnormal noise warning, and real-time monitoring of power compatibility:

[0184] Based on the collected real-time motor operating parameters, control action types, environmental parameters, and calculated power requirements Power compatibility A composite algorithm integrating tensor autoencoder and gradient boosting tree (XGBoost) is designed to achieve real-time monitoring of power fit, early prediction of insufficient power, abnormal noise warning, and preliminary differentiation of fault types. The specific process is as follows:

[0185] Feature extraction: This involves extracting features from preprocessed multi-source data (…). Type of manipulation action, sensor operating status, frequency of abnormal noise Construct it as a third-order tensor, where the first dimension is the variable dimension (14 variables), and the second dimension is the spatial location point (…). The number of sensor sampling points is the number of each sensor, consistent with the actual number of sensor groups installed. The third dimension is the time step (10 consecutive sampling points). The sampling interval (corresponding to a 100Hz sampling frequency) is used to encode the tensor using a tensor autoencoder. This process extracts low-dimensional latent space feature vectors representing motor operating status, control action type, environmental state, power adaptability, and fault characteristics. The dimension is set to 25 to capture the spatiotemporal coupling characteristics of multiple variables, actions, and environments, avoiding the loss of key information and redundant interference. The total loss function of the tensor autoencoder is defined as:

[0186] ;

[0187] in, : The total loss function of the tensor autoencoder, used to measure the fit of the model's encoding and decoding;

[0188] The squared term of the tensor reconstruction error characterizes the deviation between the input tensor and the reconstructed tensor.

[0189] : Loss function balancing coefficient, used to adjust the weight ratio of reconstruction error and KL divergence;

[0190] The Kullback-Leibler divergence (relative entropy) between two probability distributions is used to measure the difference between the posterior and prior distributions of latent variables.

[0191] Latent variables Regarding input tensors The posterior distribution follows a mean of . variance is The normal distribution;

[0192] Mean of the posterior distribution of latent variables

[0193] : Variance of the posterior distribution of latent variables;

[0194] Latent variables The prior distribution of is a standard normal distribution;

[0195] The input model consists of third-order tensor samples constructed from multi-source real-time running data;

[0196] : Low-dimensional latent space feature vectors extracted by tensor autoencoders.

[0197] Model Training: Historical database data is divided into training (70%), validation (20%), and test (10%) sets. Training is categorized by manipulation action type, environment type, and fault type. Supplementary training includes extreme operating conditions, extreme environments, and fault data. The extracted low-dimensional latent space feature vectors (25 dimensions) are input into the XGBoost model, using "power bias" as the basis for the training. ( "Absolute deviation between the actual output power of the motor and the required power" and "Power compatibility" "Fault type label" The model is trained using three labels: 0 = no fault, 1 = motor performance degradation, 2 = transmission component wear, and 3 = sensor failure. The model learns the mapping relationship between these features and power deviation, power fit, and fault type. Model parameters (learning rate, tree depth, number of leaf nodes) are optimized through grid search to keep the model's prediction error within 4% and the fault type identification accuracy ≥95%. A new false alarm correction module is added. By analyzing historical warning data, a dynamic warning threshold correction model is established (dynamically adjusting the threshold based on environmental parameters and operation action type), combined with the power fit change rate. To avoid false alarms caused by data fluctuations and environmental interference, a new extreme environment prediction sub-model has been added, which is specifically trained for high and low temperature and strong vibration environments to improve the prediction accuracy under extreme conditions.

[0198] Real-time power fit monitoring, power deficiency prediction, and extreme environment early warning: The low-dimensional latent space feature vector extracted in real time is input into the trained XGBoost model, and the power fit is output in real time. It monitors whether the motor power is sufficient to support the current operation; and predicts the power deviation in the next 50-100ms. (Model-predicted motor power deviation within the next 50-100ms), if (Power deviation threshold) or If so, it is determined that there is a risk of insufficient power, triggering an abnormal noise warning; if and If the power adaptability is within acceptable limits, it is considered a normal operating condition and no warning is triggered. For extreme operating conditions and environments, the extreme operating condition adaptation model and the extreme environment prediction sub-model are invoked to improve prediction accuracy. Simultaneously, if environmental parameters reach the extreme environment judgment threshold, an extreme environment warning is triggered, prompting operators to monitor the system status. If the power adaptability is within acceptable limits... Continued decline and Insufficient power triggers an enhanced warning.

[0199] Abnormal Noise Warning Classification and Preliminary Distinction of Fault Type: Based on Power Deviation Power compatibility Power adaptability change rate Abnormal noise detection threshold Abnormal noise frequency Based on environmental parameters, the early warning system is divided into three levels, and the fault types are initially distinguished:

[0200] Level 1 Warning: ,and The system displayed a message stating "Potential power insufficient, power adaptability low, abnormal noise imminent," triggering a minor adjustment. Initial assessment indicated no obvious fault, and the issue was most likely caused by load fluctuations.

[0201] Level 2 warning: 1.2 < ≤1.5 ,and The system displays a message indicating "High risk of insufficient power, inadequate power compatibility, and abnormal noise is about to occur," triggering moderate adjustment. Preliminary assessment suggests that there may be slight sensor error or slight performance degradation of the motor.

[0202] Level 3 Warning: >1.5 , The system displays the message "Insufficient power has occurred; the power is insufficient to support the current operation, resulting in abnormal noise," triggering a severe adjustment and locating the motor of the actuator with insufficient power; combined with the frequency of the abnormal noise... Based on motor operating parameters and sensor operating status, the fault type can be preliminarily distinguished: if the frequency of abnormal noise is positively correlated with the motor speed and the motor winding temperature is normal, it is determined to be motor performance degradation; if the frequency of abnormal noise is positively correlated with the displacement of the actuator, it is determined to be wear of transmission components; if the sensor operating status is abnormal and the data fluctuates greatly, it is determined to be sensor fault.

[0203] Multi-modal power adaptive adjustment with multi-action adaptation:

[0204] Based on the prediction results, warning level, type of control action, environmental parameters, and preliminary fault type, the corresponding power adjustment mode is activated. A three-modal adjustment mechanism is employed: "action-adaptive base power + dynamic compensation power + extreme environment correction power." Combined with variable frequency volumetric speed control technology, this achieves adaptive adjustment of motor power, ensuring... and The deviation was always controlled within within, To suppress abnormal noises and ensure sufficient motor power to support the current operation, the following adjustment procedure is followed:

[0205] Base power setting: Calculated The motor's base power is set based on the current operating action type and environmental parameters, using the following formula:

[0206] (Pitch and yaw motions are taken as 0.05, roll and combined motions are taken as 0.08).

[0207] in, : The base power set for the motor at any given time;

[0208] : Real-time power demand of the motor.

[0209] It reserves corresponding power redundancy to avoid power shortage caused by instantaneous load fluctuations during operation switching; at the same time, based on variable frequency volumetric speed regulation technology, the power matching efficiency is improved to over 90%, reducing energy consumption; a new dynamic correction logic for basic power is added, based on the motor winding temperature. Motor running time Dynamic correction of environmental parameters: when When the temperature exceeds 85℃, appropriately reduce the basic power redundancy (to 0.03~0.05) to avoid motor overheating; when in extreme environments, appropriately increase the basic power redundancy (to 0.08~0.1) to improve system stability; when the motor runs for a long time... For motors operating for more than 1000 hours, increase the basic power redundancy (to 0.06~0.09) to compensate for motor performance degradation.

[0210] Dynamic compensation power and extreme environment correction power calculation: based on warning level and power deviation. Power compatibility Power adaptability change rate Based on environmental parameters, calculate the dynamic compensation power. Power correction for extreme environments The PID optimization algorithm (introducing fuzzy control to improve regulation robustness) is used for regulation. The input of the PID controller is the power deviation. Power compatibility deviation and environmental deviation (Deviation between actual and standard environments), the output is the total compensation power adjustment, and the corrected PID control formula is:

[0211] ;

[0212] in, : The total compensation power adjustment of the motor at any given time;

[0213] : Dynamic power compensation of the motor at all times;

[0214] : Power correction in extreme environments;

[0215] The proportional gain of the PID controller is dynamically adjusted according to the warning level and environmental type.

[0216] The integral coefficient of the PID controller is dynamically adjusted according to the warning level and environmental type.

[0217] The derivative coefficient of the PID controller is dynamically adjusted according to the warning level and environmental type.

[0218] : The absolute deviation between the actual output power of the motor and the required power at any given time;

[0219] : Momentary power adaptation deviation ;

[0220] : The deviation between actual environmental parameters and standard environmental parameters at any given time;

[0221] Integration time variable;

[0222] Integral term The time integral of the deviation is used to eliminate the steady-state error of the system.

[0223] Differential term The time derivative of the deviation is used to predict the trend of deviation changes and improve the system response speed.

[0224] PID parameters are dynamically adjusted based on the warning level and environment type:

[0225] Level 1 Warning (Normal Environment): =0.3 ， = 0.05 ， =0.1 ， =0.05× ( + ))， =0;

[0226] Level 2 Warning (Normal Environment): =0.5 ， =0.1 ， =0.2 ， =0.1× ( + )， =0.02 × ;

[0227] Level 3 Warning (Normal Environment): =0.8 ， =0.2 ， =0.3 ， =0.15× ( + )， =0.03× ;

[0228] Extreme Environment (Any Warning Level): Based on the parameters of the corresponding warning level, 、 、 Both increased by 20%. Increased to 0.05× To ensure adjustment accuracy under extreme environments.

[0229] Actual power output adjustment: Real-time output power of the motor The motor voltage is adjusted by the controller. With current This enables real-time power adjustment; a new motion switching prediction module has been added, which analyzes the changing trends of control motion recognition data (such as the rate of change of angular velocity and angular acceleration) to predict motion switching in advance (such as pitch to roll, changes in the amplitude of linked motions), and adjusts the power adjustment response speed 50ms in advance. PID parameters have been optimized, and an instantaneous power buffer mechanism (power change rate limit) has been introduced. (Maximum allowable change in power per unit time) to avoid instantaneous power shortage and power surges during action switching, and to mitigate the risk of action switching lag; a predictive model reference control method is introduced, which predicts the next operating state based on the motor response prediction model, and fine-tunes the control command amplitude according to the error correction coefficient within the control cycle to ensure the smoothness of adjustment, avoid secondary abnormal noises caused by power surges, and ensure power adaptability during operation action switching; targeted adjustments are made for the initially determined fault type: if it is determined to be a sensor fault, adjustment is made using backup sensor data; if it is determined to be a motor performance degradation, the compensation power is appropriately increased; if it is determined to be wear of transmission components, the power adjustment frequency is reduced to reduce wear of transmission components.

[0230] Adjustment constraints and full-process motor protection: To avoid motor overload, the maximum output power of the motor is set. (Maximum allowable output power of the motor), if Then Forced to be set It also issues a "motor overload warning" and activates a load diversion mechanism to coordinate other actuator motors for the corresponding operation to assist in bearing the load, preventing damage from prolonged overload of a single motor; and monitors the motor winding temperature in real time. Motor operating current ,when When the temperature exceeds 90℃, activate motor cooling measures (such as forced air cooling or cooling spray) and reduce dynamic compensation power. If the temperature continues to rise to 95℃, issue a "motor overheat warning" and stop power regulation of that motor, switching to a backup motor (if applicable); limit the frequency of motor power regulation (every...). Adjust at most once (avoid frequent adjustments that could increase motor wear and tear, and also consider the duration of motor operation). For more than 1000 hours, reduce the adjustment frequency (per... (Maximum adjustment once) to compensate for motor performance degradation; Added motor performance degradation monitoring; when motor efficiency... η If the value remains below 0.8, a "motor performance degradation warning" will be issued, prompting the operator to perform maintenance; when wear characteristics of transmission components (friction load) are detected... (If the temperature continues to rise), a "transmission component wear warning" will be issued, prompting operators to check and replace the affected components.

[0231] Closed-loop feedback regulation and model updating:

[0232] A closed-loop control system is constructed, consisting of "real-time power monitoring, prediction, adjustment, feedback, risk prevention and control, and fault handling." This system combines various control action types and environmental parameters to monitor the adjustment effect, motor operating status, sensor status, and various risk points in real time. Model parameters are then corrected to achieve continuous optimization, accurately locate faults, and form a closed-loop risk handling mechanism. The specific process is as follows:

[0233] Monitoring of adjustment effect, risk status and fault status: Real-time acquisition of motor power after adjustment. Intensity of abnormal noise Abnormal noise frequency Power deviation Power compatibility Power adaptability change rate Motor winding temperature Motor running time Based on sensor operating status and environmental parameters, combined with the current type of manipulation and preliminary fault type, determine the adjustment effect, risk status, and fault status:

[0234] like ≤ 、 ≥ 、 < 、 If the temperature is ≤85℃, the sensor is operating normally, and the environmental parameters are within the normal range, it indicates that the adjustment is effective and risk-free. The current adjustment parameters should be maintained.

[0235] like > 、 < or ≥ This indicates that the adjustment effect is not good; if If the temperature exceeds 85℃, the motor efficiency remains below 0.8, and the friction of the transmission components continues to increase, it indicates a risk of motor overheating, performance degradation, and wear of the transmission components. If the sensor operates abnormally, it indicates a risk of sensor failure. If the environmental parameters are in extreme ranges, it indicates an extreme environmental risk. All of these situations require parameter correction and risk management.

[0236] Parameter correction and risk closed-loop management: Based on the deviation of the adjustment effect, the current type of operation, environmental parameters, risk status, and initial fault type, dynamically correct the PID controller parameters. 、 、 The coupling model parameters and extreme environment correction coefficients are given by the following formula:

[0237] ;

[0238] ;

[0239] ;

[0240] in, The corrected proportional, integral, and derivative coefficients of the PID controller;

[0241] After power adjustment, The absolute deviation between the actual output power of the motor and the required power at any given time;

[0242] After power adjustment, Real-time power adaptation of the motor;

[0243] : The deviation between actual environmental parameters and standard environmental parameters at any given time;

[0244] : The maximum deviation value of the preset extreme environment parameters.

[0245] At the same time, the efficiency parameters in the coupled model are corrected. The model is configured with load weights and environmental correction coefficients for different maneuvers to ensure consistency between the model and actual operating conditions, maneuver types, and environmental states; closed-loop management is implemented for various risks.

[0246] Motor overheating risk: Increase cooling measures and adjust the basic power redundancy. If the temperature continues to rise, switch to the backup motor.

[0247] Sensor malfunction risk: Switch to the backup sensor, issue a sensor maintenance prompt, and correct the sensor error model. If the backup sensor is also malfunctioning, initiate correlation data fitting adjustment.

[0248] Extreme environment risks: Optimize the correction coefficient for extreme environments, increase the basic power redundancy, reduce the power regulation frequency, and ensure system stability;

[0249] Risks of motor performance degradation and wear of transmission components: Issue maintenance warnings and adjust parameters to reduce motor losses and extend service life.

[0250] Online model updates, generalized optimization, and precise fault localization: An incremental training strategy is adopted, categorized by manipulation action type, environment type, and fault type, and adjusted real-time data (including...) are then used. 、 、 、 (Including multi-source operating parameters, operation action types, environmental parameters, risk handling records, and fault types) added to the historical database, every 100 sampling cycles ( Sampling period 100× =1s), perform a fine-tuning iteration on the tensor autoencoder and XGBoost models, update the model parameters, and the corrected model parameter update formula is:

[0251] ;

[0252] in, The updated model parameter vector after incremental training;

[0253] : The original model parameter vector before the update;

[0254] The learning rate for incremental training of the model controls the step size for updating the model parameters in each iteration.

[0255] The gradient operator is used to calculate the partial derivatives of the loss function with respect to the model parameters.

[0256] The gradient of the loss function is the set of partial derivatives of the loss function with respect to each parameter, calculated based on the model parameters before the update.

[0257] When a control performance indicator deviates from a set threshold, the model migration module is triggered to switch to a preset backup model, ensuring the stability of prediction, power monitoring, and adjustment. A new generalized model optimization module has been added, which analyzes the operating data of different simulator models, extracts common features, optimizes the model's general parameters, and includes parameter templates for multiple simulator models to reduce the difficulty of adapting to different simulator models and improve the versatility of the solution. It also supports one-click import and export of parameters to further reduce adaptation costs and avoid the risk of insufficient versatility. A new fault accurate location module has been added, which combines the fault type labels, real-time operating parameters, and abnormal noise characteristics output by the model with historical fault data to accurately locate the fault location (specific motor, sensor, transmission component) and the degree of fault, and outputs targeted maintenance suggestions (such as "the performance of the No. 1 actuator motor has deteriorated, it is recommended to repair the winding" and "the No. 3 angle sensor is faulty, it is recommended to replace it"). It also refers to troubleshooting methods such as cross-connection to improve fault location efficiency, reduce manual troubleshooting costs, and solve the problem of inaccurate fault location.

[0258] Fault handling verification: If the abnormal noise persists even after intensive adjustments ( continued If the abnormal noise persists beyond the threshold for a certain duration (or the power compatibility remains below 0.7), and the current operation is normal, the sensors are functioning normally, and the environmental parameters are normal, then based on the accurate fault location results, the system will prompt the operator to inspect the faulty part. After the inspection is completed, the system will automatically restart the control process to verify the fault handling effect and ensure that the abnormal noise is eliminated and the power compatibility is normal.

[0259] System terminated:

[0260] When the pilot ceases operational actions, the simulator returns to its initial state, the motors return to standby mode, and the power output decreases to the corrected standby power formula:

[0261] ;

[0262] in, : The output power of the motor in standby mode, the basic power maintained by the motor when the simulator is not operating.

[0263] The system stops power regulation and power adaptability monitoring, and saves the data of this operation (including multi-source data, operation action type, environmental parameters, power regulation process, power adaptability change, early warning information, regulation effect, risk handling record, fault type and location result) to the historical database for subsequent model optimization, fault tracing, generalized parameter iteration and fault diagnosis model training.

[0264] For the purpose of simplicity, the method steps disclosed in the above embodiments are described as a series of actions. However, those skilled in the art should understand that the embodiments of the present invention are not limited to the described order of actions, because according to the embodiments of the present invention, some steps can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are all preferred embodiments, and the actions involved are not necessarily essential to the embodiments of the present invention.

[0265] like Figure 2 As shown, the present invention also provides a simulator actuator motor power control system, comprising:

[0266] The system initialization and parameter calibration module 201 is configured to set basic parameters and establish a multi-dimensional sensor error correction model.

[0267] The multi-source data real-time acquisition and preprocessing module 202 is configured to acquire multiple types of operating parameters in real time, select key data, perform real-time error calibration based on the multi-dimensional sensor error correction model, and obtain preprocessed real-time operating data.

[0268] The coupling model construction and power demand and adaptability calculation module 203 is configured to construct a coupled mathematical model of multiple operation actions-load-power based on the real-time operation data, calculate the real-time power demand of the actuator motor, and calculate the power adaptability of the actuator motor based on the real-time output power of the actuator motor and the real-time power demand.

[0269] The power insufficiency prediction, abnormal noise warning and power adaptability real-time monitoring module 204 is configured to obtain the prediction result of the power insufficiency risk in the future period through the prediction model based on the real-time operation data, the real-time power demand and the power adaptability, and output the abnormal noise warning according to the prediction result. At the same time, it monitors the power adaptability to obtain the judgment result of whether the real-time power of the actuator motor is sufficient to support the current operation.

[0270] The multimodal power adaptive adjustment and closed-loop feedback module 205 is configured to adjust the output power of the actuator motor using a multimodal adjustment mechanism based on the prediction result and the judgment result, and to monitor the power adaptability and abnormal noise intensity after adjustment, and to correct the parameters of the control model based on the monitoring result, wherein the control model includes the coupled mathematical model and the prediction model.

[0271] It is worth noting that although only some basic functional modules are disclosed in the embodiments of this invention, it does not mean that the composition of this system is limited to the above-mentioned basic functional modules. On the contrary, what this embodiment intends to express is that, based on the above-mentioned basic functional modules, those skilled in the art can arbitrarily add one or more functional modules in combination with existing technology to form an infinite number of embodiments or technical solutions. That is to say, this system is open rather than closed. The fact that this embodiment only discloses a few basic functional modules should not be considered as the scope of protection of the claims of this invention being limited to the disclosed basic functional modules. At the same time, for the convenience of description, the above device is described separately according to its functions as various units and modules. Of course, in implementing this invention, the functions of each unit and module can be implemented in one or more software and / or hardware.

[0272] like Figure 3 As shown, the present invention also provides an electronic device, including: a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus; the memory stores a computer program, and when the computer program is executed by the processor, the processor performs the steps of a method for controlling the power of a simulator actuator motor.

[0273] Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Figure 3 The structure shown in this embodiment of the invention includes an electronic device comprising one or more processors 710 and a memory 720; the processors 710 in this electronic device may be one or more. Figure 3 Taking a processor 710 as an example; a memory 720 is used to store one or more programs; the one or more programs are executed by the one or more processors 710, so that the one or more processors 710 implement a simulator actuator motor power control method as described in any one of the embodiments of the present invention.

[0274] The electronic device may also include an input device 730 and an output device 740.

[0275] The processor 710, memory 720, input device 730, and output device 740 in this electronic device can be connected via a bus or other means. Figure 3 Taking the example of a connection between China and Israel via a bus.

[0276] The memory 720 in this electronic device serves as a computer-readable storage medium, capable of storing one or more programs. These programs can be software programs, computer-executable programs, or modules, such as the program instructions / modules corresponding to the simulator actuator motor power control method provided in this embodiment of the invention. The processor 710 executes various functional applications and data processing of the electronic device by running the software programs, instructions, and modules stored in the memory 720, thereby implementing the simulator actuator motor power control method described in the above embodiment.

[0277] The memory 720 may include a program storage area and a data storage area. The program storage area may store the operating system and applications required for at least one function; the data storage area may store data created based on the use of the electronic device. Furthermore, the memory 720 may include high-speed random access memory and may also include non-volatile memory, such as at least one disk storage device, flash memory device, or other non-volatile solid-state storage device. In some instances, the memory 720 may further include memory remotely located relative to the processor 710, which can be connected to the device via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0278] Input device 730 can be used to receive input digital or character information, and to generate key signal inputs related to user settings and function control of the electronic device. Output device 740 may include display devices such as a display screen.

[0279] The present invention also provides a computer-readable storage medium storing a computer program executable by an electronic device, which, when run on the electronic device, causes the electronic device to perform the steps of a method for controlling the power of a simulator actuator motor.

[0280] Specifically, the computer storage medium in this embodiment of the invention can be any combination of one or more computer-readable media. The computer-readable medium can be a computer-readable signal medium or a computer-readable storage medium. For example, a computer-readable storage medium can be—but is not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of computer-readable storage media (a non-exhaustive list) include: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this embodiment, the computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0281] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for controlling the power of a simulator actuator motor, characterized in that, include: Set basic parameters and establish a multi-dimensional sensor error correction model; Multiple types of operating parameters are collected in real time, key data are selected, and real-time error calibration is performed based on the multi-dimensional sensor error correction model to obtain preprocessed real-time operating data. Based on the real-time operation data, a coupled mathematical model of multiple control actions-load-power is constructed to calculate the real-time power requirement of the actuator motor, and the power adaptability of the actuator motor is calculated based on the real-time output power of the actuator motor and the real-time power requirement. Based on the real-time operating data, the real-time power demand, and the power adaptability, a prediction result of the power insufficiency risk in the future period is obtained through the prediction model, and an abnormal noise warning is output according to the prediction result. At the same time, the power adaptability is monitored to obtain a judgment result on whether the real-time power of the actuator motor is sufficient to support the current operation. Based on the prediction results and the judgment results, a multi-modal adjustment mechanism is used to adjust the output power of the actuator motor, and the power adaptability and abnormal noise intensity after adjustment are monitored. The parameters of the control model are corrected based on the monitoring results. The control model includes the coupled mathematical model and the prediction model. Among them, the preprocessed real-time running data is constructed into a tensor, and low-dimensional latent space feature vectors are extracted through tensor autoencoder; The low-dimensional latent space feature vector is input into the gradient boosting tree model trained by classifying and training the historical database according to the type of manipulation action, the type of environment, and the type of fault, to obtain the predicted value of power deviation, power fit, and fault type label. Based on the predicted power deviation, power adaptability, power adaptability change rate, real-time abnormal noise intensity, abnormal noise frequency and environmental parameters, the abnormal noise warning is divided into multiple levels, and based on the correlation between abnormal noise frequency and motor speed, the correlation between abnormal noise frequency and actuator displacement, and sensor working status, preliminary fault type differentiation results are obtained. By analyzing historical early warning data, a dynamic correction model for early warning thresholds is established. The thresholds are dynamically adjusted according to environmental parameters and the type of manipulation action. When the power adaptability is lower than the power adaptability threshold, a power insufficiency warning is triggered. When the power adaptability change rate is lower than the set threshold, a power insufficiency aggravation warning is triggered.

2. The method for controlling the power of a simulator actuator motor according to claim 1, characterized in that, Set basic parameters and establish a multi-dimensional sensor error correction model, which further includes: The basic parameters include at least: Rated power of the actuator motor, angle range of various operating actions, load fluctuation threshold, power deviation threshold, abnormal noise detection threshold, power adaptability threshold, and extreme environment judgment threshold; A configurable parameter interface is set up to configure parameters such as actuator stiffness coefficient, transmission mechanism friction coefficient, actuator motor efficiency benchmark value, transmission efficiency benchmark value, and actuator lever arm corresponding to various operating actions according to the simulator model. Parameter templates for multiple simulator models are also built-in. By using temperature compensation algorithm, vibration filtering algorithm, and drift calibration algorithm, the sensor error coefficients under different temperature, vibration, and operating length conditions are calibrated, and the multi-dimensional sensor error correction model is established. A historical database is built to store historical operating data, including motor power, abnormal noise intensity, power adaptability data, extreme operating condition data, and fault data for various operating actions, different angles, different loads, and different temperatures.

3. The method for controlling the power of a simulator actuator motor according to claim 1, characterized in that, Real-time acquisition of multiple types of operating parameters, selection of key data, and real-time error calibration based on the multi-dimensional sensor error correction model yield preprocessed real-time operating data, further including: Real-time data acquisition at a set sampling frequency: Real-time current, real-time voltage, real-time speed, winding temperature, and running time of the actuator motor; real-time pitch angle, roll angle, yaw angle and corresponding angular velocity and angular acceleration of the simulator cockpit; real-time load of the cockpit, ambient temperature, ambient vibration intensity, and real-time abnormal noise intensity and frequency. Based on a dual backup configuration combining a primary sensor and a backup sensor, when the primary sensor fails, it automatically switches to the backup sensor data; when the backup sensor also fails, it supplements the missing data by fitting correlated data. The collected data is denoised and abnormal data is removed by time-domain averaging, sliding window integration and median filtering, and real-time error calibration is performed by the multi-dimensional sensor error correction model. The calibrated data is standardized to obtain preprocessed real-time operating data.

4. The method for controlling the power of a simulator actuator motor according to claim 1, characterized in that, Based on the real-time operating data, a coupled mathematical model of multiple control actions, load, and power is constructed to calculate the real-time power demand of the actuator motor. Furthermore, the power adaptability of the actuator motor is calculated based on its real-time output power and the real-time power demand. The load composition of various control actions is analyzed. The total load borne by the actuator is calculated by the cockpit gravity load, the inertial load of the control action, and the friction load of the transmission mechanism. Among them, the gravity load and inertial load are calculated differently according to the pitch, roll, yaw and linkage action types, and an environmental vibration correction coefficient is introduced. For linkage actions, the total load is the weighted sum of the loads of each individual action, and an action switching correction coefficient is introduced. Based on the total load, the real-time speed of the actuator cylinder, the real-time operating efficiency of the actuator cylinder motor, and the real-time operating efficiency of the transmission mechanism, the real-time power requirement of the actuator cylinder motor is calculated. The real-time speed of the actuator cylinder is obtained by differentiating the displacement of the actuator cylinder with respect to time. Based on the real-time operating voltage, real-time operating current and real-time operating efficiency of the actuator motor, the real-time output power of the actuator motor is calculated, and the ratio of the real-time output power to the real-time power demand is used as the power adaptability of the actuator motor. At the same time, the power adaptability change rate is calculated. The real-time operating efficiency and the real-time operating efficiency of the transmission mechanism are dynamically corrected based on the operating time of the actuator motor, ambient temperature, winding temperature and ambient vibration intensity. The potential correlation features between load and power in multi-source coupled data are extracted by tensor autoencoder, and the model parameters of similar working conditions in historical data are transferred to extreme working conditions by transfer learning algorithm.

5. The method for power regulation of a simulator actuator motor according to claim 1, characterized in that, Based on the prediction result and the determination result, a multi-modal adjustment mechanism is used to adjust the output power of the actuator motor, further including: Based on the real-time power requirements, the base power of the actuator motor is set according to the current operation type and environmental parameters, and power redundancy is reserved in the base power. The base power redundancy value is dynamically adjusted according to the winding temperature, running time and environmental parameters of the actuator motor. Based on the warning level, the predicted value of power deviation, the power adaptability, the power adaptability change rate and environmental parameters, the dynamic compensation power and the extreme environment correction power are calculated by the PID control algorithm to obtain the total compensation power adjustment. The control parameters of the PID controller are dynamically adjusted according to the warning level and the environmental type. Based on the base power and the total compensation power adjustment, power output is achieved by adjusting the voltage and current of the actuator motor, limiting the power change rate, predicting action switching by analyzing the changing trend of the control action identification data, and adjusting the control parameters in advance. Monitor the winding temperature and operating current of the actuator motor. When the winding temperature exceeds the temperature threshold, activate cooling measures or switch to the standby motor. When the real-time output power of the actuator motor exceeds the protection multiple of the rated power of the actuator motor, activate the load shunting mechanism and limit the power regulation frequency of the actuator motor.

6. The method for controlling the power of a simulator actuator motor according to claim 5, characterized in that, The adjusted power fit and abnormal noise intensity are monitored, and the parameters of the control model are corrected based on the monitoring results, further including: Collect the adjusted actuator motor power, abnormal noise intensity, abnormal noise frequency, predicted power deviation, power adaptability, power adaptability change rate, and winding temperature. Combine this with the current operating action type and environmental parameters to determine the adjustment effect and risk status. When the adjustment effect is not good or there is a risk, the control parameters of the PID controller are dynamically corrected according to the adjustment effect deviation, the current operation action type, environmental parameters and risk status. The efficiency parameters, load weights of different operation actions and environmental correction coefficients in the coupled mathematical model are corrected, and the risk is handled in a closed loop. An incremental training strategy is adopted, which classifies the data according to the type of manipulation action, environment type, and fault type. The adjusted real-time data is added to the historical database to iteratively update the prediction model. When the control performance index deviates from the set threshold, the model is switched to the backup model. By combining fault type labels, real-time operating parameters, and abnormal noise characteristics, and comparing historical fault data, the system outputs the location results of the fault location and fault severity, and provides maintenance suggestions.

7. A power control system for a simulator actuator motor, characterized in that, include: The system initialization and parameter calibration module is configured to set basic parameters and establish a multi-dimensional sensor error correction model. The multi-source data real-time acquisition and preprocessing module is configured to acquire multiple types of operating parameters in real time, select key data, perform real-time error calibration based on the multi-dimensional sensor error correction model, and obtain preprocessed real-time operating data. The coupling model construction and power demand and adaptability calculation module is configured to construct a coupled mathematical model of multiple operation actions-load-power based on the real-time operation data, calculate the real-time power demand of the actuator motor, and calculate the power adaptability of the actuator motor based on the real-time output power of the actuator motor and the real-time power demand. The power insufficiency prediction, abnormal noise warning and power adaptability real-time monitoring module is configured to obtain the prediction result of the power insufficiency risk in the future period through the prediction model based on the real-time operation data, the real-time power demand and the power adaptability, and output the abnormal noise warning according to the prediction result. At the same time, it monitors the power adaptability to obtain the judgment result of whether the real-time power of the actuator motor is sufficient to support the current operation. The multimodal power adaptive adjustment and closed-loop feedback module is configured to adjust the output power of the actuator motor using a multimodal adjustment mechanism based on the prediction result and the judgment result, and to monitor the power adaptability and abnormal noise intensity after adjustment, and to correct the parameters of the control model based on the monitoring result, wherein the control model includes the coupled mathematical model and the prediction model; Among them, the preprocessed real-time running data is constructed into a tensor, and low-dimensional latent space feature vectors are extracted through tensor autoencoder; The low-dimensional latent space feature vector is input into the gradient boosting tree model trained by classifying and training the historical database according to the type of manipulation action, the type of environment, and the type of fault, to obtain the predicted value of power deviation, power fit, and fault type label. Based on the predicted power deviation, power adaptability, power adaptability change rate, real-time abnormal noise intensity, abnormal noise frequency and environmental parameters, the abnormal noise warning is divided into multiple levels, and based on the correlation between abnormal noise frequency and motor speed, the correlation between abnormal noise frequency and actuator displacement, and sensor working status, preliminary fault type differentiation results are obtained. By analyzing historical early warning data, a dynamic correction model for early warning thresholds is established. The thresholds are dynamically adjusted according to environmental parameters and the type of manipulation action. When the power adaptability is lower than the power adaptability threshold, a power insufficiency warning is triggered. When the power adaptability change rate is lower than the set threshold, a power insufficiency aggravation warning is triggered.

8. An electronic device, characterized in that, include: The system includes a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other via the communication bus; the memory stores a computer program, which, when executed by the processor, causes the processor to perform the steps of the method according to any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, It stores a computer program executable by an electronic device, which, when run on the electronic device, causes the electronic device to perform the steps of the method according to any one of claims 1 to 6.

Images