Force sense calibration method for aircraft throttle lever

Abstract

The invention belongs to the technical field of aircraft maintenance and debugging, particularly relates to a force sensing calibration method for an aircraft throttle lever, and aims to solve the problems that an existing aircraft throttle lever force sensing calibration technology depends on manpower, and is low in efficiency, insufficient in precision and the like. The method comprises the following steps: generating a continuous force-position curve; identifying a positioning point in the force-position curve, and dividing the whole stroke into a positioning section and a smooth stroke section; calculating basic friction force and positioning retention force, and calculating deviation values of the basic friction force and the positioning retention force relative to target values; the method comprises the following steps: applying a test adjustment amount to an adjustment mechanism, measuring a generated cross variation, and constructing an influence coefficient matrix; and calculating an inverse matrix of the influence coefficient matrix, calculating a target physical adjustment amount in combination with the deviation amount, and driving the friction adjustment mechanism and the positioning adjustment mechanism to act simultaneously. According to the invention, the basic friction force and the positioning retention force are rapidly, accurately and synchronously adjusted, so that the calibration efficiency and precision are remarkably improved.

Description

Technical Field

[0001] This invention belongs to the field of aircraft maintenance and debugging technology, and specifically relates to a force-sensing calibration method for an aircraft throttle lever. Background Technology

[0002] In modern civil aircraft, the throttle lever's tactile feedback is a critical parameter for ensuring precise and consistent pilot control, thereby guaranteeing flight safety. Manufacturers have established rigorous force calibration procedures for this purpose. Currently, the traditional calibration methods commonly used in the industry mainly rely on manual operation, which has many drawbacks: First, relying on technicians to manually measure and judge using handheld force gauges results in reading errors due to parallax and uneven force application, leading to poor repeatability of measurement results; second, the adjustment process is cumbersome and inefficient, requiring manual rotation of adjusting nuts and screws, and multiple iterations of "measurement-adjustment-remeasurement" to approximate the target value, which is time-consuming and labor-intensive; third, for measuring the breakthrough force at the positioning point, it relies on the technician's subjective feeling of a "sudden increase" in force, which is inaccurate to identify and difficult to reproduce; in addition, the calibration process lacks complete digital recording and traceability, usually relying on manual filling of forms, which cannot record complete force-displacement curves and adjustment processes, hindering quality control and in-depth analysis; finally, the accuracy and efficiency of the entire calibration process are highly dependent on the experience of technicians, resulting in high training costs and difficulty in ensuring consistency, while existing methods only focus on achieving the target in a single calibration, lacking the ability to track and predict the long-term performance evolution trend of the throttle lever force sensor unit.

[0003] Therefore, there is an urgent need in this field for a throttle lever force calibration scheme that can achieve high precision, high efficiency, automation, and complete data traceability and health prediction capabilities, in order to overcome the above-mentioned technical defects of relying on manual labor, low efficiency, insufficient accuracy, lack of traceability, and inability to achieve preventive maintenance. Summary of the Invention

[0004] To address the aforementioned problems in existing technologies, namely the reliance on manual labor, low efficiency, and insufficient accuracy in current aircraft throttle lever force-sensing calibration techniques, this invention provides a force-sensing calibration method for aircraft throttle levers, comprising:

[0005] Simultaneously acquire force and position signals of the throttle lever during its full stroke operation to generate a continuous force-position curve; Based on the preset feature template, the positioning point abrupt change features in the force-position curve are identified, and the entire stroke of the throttle lever is divided into positioning segments containing positioning point abrupt change features and smooth stroke segments located between positioning points. The average force signal within the smooth travel segment is calculated as the basic friction force, and the peak force signal within the positioning segment is extracted as the positioning holding force. The deviations of the basic friction force and the positioning holding force from their respective preset target values ​​are calculated. A preset amplitude test adjustment amount is applied to the friction adjustment mechanism and the positioning adjustment mechanism of the throttle lever respectively, and the cross change in the basic friction force and the positioning holding force caused by the test action of the two adjustment mechanisms is measured. An influence coefficient matrix is ​​constructed using sensitivity coefficients; the sensitivity coefficients are calculated based on the ratio of the cross-variance to the experimental adjustment. The inverse matrix of the influence coefficient matrix is ​​calculated, and the deviation is operated on with the inverse matrix to solve for the target physical adjustment amount used to eliminate the deviation, and the friction adjustment mechanism and the positioning adjustment mechanism are driven to operate simultaneously.

[0006] Furthermore, the steps for generating a continuous force-position curve specifically include: Force sensor data and position encoder data are acquired synchronously at a preset high sampling frequency and time-aligned. An adaptive filtering algorithm is applied to process the original force signal to filter out mechanical vibration noise and operational jitter noise; The denoised force signal is mapped to the position coordinate axis to construct a continuous curve of force changing with position. The first and second derivative curves of the continuous curve with respect to position are calculated, where the first derivative curve is used to help identify the abrupt change rate of the force value.

[0007] Furthermore, the steps of dividing the throttle lever's full travel into a positioning segment and a smooth travel segment specifically include: Establish a key feature template that includes the range of force value peak amplitude, the span of wave width position, and the shape characteristics of derivative; Scan the first derivative curve to identify local extrema and zero intersections, and extract candidate mutation regions; The candidate mutation regions and feature templates are cross-correlated and normalized to calculate morphological similarity. Valid localization points are then confirmed by combining the travel position constraints. The positioning segment is defined as the position confirmed as a valid positioning point, extending a preset width to both sides. The continuous travel outside the positioning segment is defined as the smooth travel segment.

[0008] Furthermore, the specific steps for constructing the influence coefficient matrix include: The changes in basic friction force and positioning holding force caused by the application of the test adjustment amount of the friction adjustment mechanism are obtained, and then divided by the test adjustment amount to obtain the first set of sensitivity coefficients, which are used as the first column elements of the influence coefficient matrix. The changes in basic friction force and positioning holding force caused by the application of the test adjustment amount of the positioning adjustment mechanism are obtained, and then divided by the test adjustment amount to obtain the second set of sensitivity coefficients, which are used as the second column elements of the influence coefficient matrix. A two-dimensional influence coefficient matrix is ​​established based on the elements in the first column and the elements in the second column.

[0009] Furthermore, the specific steps for calculating the target physical adjustment amount include: Calculate the pseudo-inverse matrix of the influence coefficient matrix, perform matrix multiplication on the pseudo-inverse matrix and the deviation vector composed of the basic friction force deviation and the positioning holding force deviation, and use the calculation result as the target physical adjustment vector containing the rotation adjustment amount of the friction adjustment mechanism and the rotation adjustment amount of the positioning adjustment mechanism.

[0010] Furthermore, the specific steps for simultaneously driving the friction adjustment mechanism and the positioning adjustment mechanism include: A variable gain control strategy is adopted, which converts the calculated target physical adjustment amount into the control amount of the servo motor according to different proportions based on the current force deviation amount; When the absolute value of the deviation is greater than the first preset threshold, a large adjustment is performed using the first gain coefficient. When the absolute value of the deviation is less than or equal to the first preset threshold but greater than the second preset threshold, a second gain coefficient less than the first gain coefficient is used to perform fine adjustment. After one or more adjustment operations, the force-position curve is re-acquired and the current force value is calculated. If the basic friction force or positioning holding force deviates from the preset tolerance range due to the fixed operation after adjustment, a new round of calculation and adjustment is automatically started based on the new deviation until the two force parameters are simultaneously stable within the preset tolerance range.

[0011] Furthermore, this method also includes: Extract health characteristic parameters from the calibrated force-position curve; By combining the health characteristic parameters of this calibration with historical calibration data, a time series analysis was conducted to calculate the health index, which characterizes the degree of performance degradation of the throttle lever force sensor unit. Based on the current value and trend of the health index, predictive maintenance recommendations are generated.

[0012] Furthermore, the steps for extracting the health characteristic parameters of the calibrated force-position curve specifically include: The dispersion index of the second derivative of the force-position curve within the smooth stroke section is calculated and used as a smoothness index characterizing the mechanical smoothness. Obtain the push rod process curve and the pull rod process curve respectively, and calculate the integral area of ​​the force difference between the two at the same stroke position over the stroke, which is used as the hysteresis loop parameter characterizing mechanical clearance and internal friction. The stroke is divided into several sub-intervals, and the attenuation ratio of the average force value of each sub-interval relative to the historical initial state of the throttle lever is calculated as the segment attenuation rate parameter. The peak value of the force change rate of the positioning holding force on the rising edge is calculated as a steepness parameter characterizing the roller wear state of the positioning mechanism.

[0013] Furthermore, the specific steps for calculating the health index, which characterizes the degree of performance degradation of the throttle lever force-sensing unit, include: The characteristic parameters in the historical calibration records of the aircraft were retrieved, and the time series analysis method was applied to fit the parameter evolution trend line to obtain the slope of the trend line. Retrieve the statistical distribution of characteristic parameters of the same aircraft type group, and calculate the deviation score of the current characteristic parameter relative to the group mean. Using the standard reference curve provided by the manufacturer as a reference, calculate the dynamic time bending distance between the current force-position curve and the reference curve; The slope of the trend line, the deviation score, and the dynamic time bending distance are weighted and fused to generate a normalized health index.

[0014] Furthermore, based on the current value and trend of the health index, the specific steps for generating predictive maintenance recommendations include: The extracted health feature parameters are compared with preset fault feature thresholds or historical benchmark ranges. When the smoothness index exceeds the first abnormal threshold, maintenance suggestions are generated to check the lubrication status of mechanical moving parts or the presence of foreign objects. When the hysteresis loop area parameter exceeds the second abnormal threshold, maintenance recommendations for checking transmission mechanism clearance or bearing wear are generated. When the force attenuation rate of a specific stroke sub-segment exceeds the third abnormal threshold, maintenance recommendations are generated to check the wear status of the friction pads in the corresponding area. Based on the predicted time when the health index will drop to the preset warning threshold, a preventative replacement or in-depth maintenance work order is generated.

[0015] The beneficial effects of this invention are: This method achieves objective, accurate, and repeatable measurement of throttle lever force parameters through high-precision synchronous sensing, adaptive filtering, and intelligent feature recognition. It completely eliminates errors introduced by human-specific parallax, uneven force application, and subjective judgment in traditional manual measurements, ensuring the consistency of the calibration benchmark. Furthermore, its core decoupled closed-loop control strategy, by constructing and utilizing the influence coefficient matrix between the adjustment mechanisms, can intelligently calculate the optimal coordinated adjustment amount to eliminate force deviations, driving the synchronous action of both mechanisms. This fundamentally solves the problem of repeated oscillations and inefficiency caused by mechanical coupling in traditional "trial and error" adjustments, enabling the basic friction force and the positioning point breakthrough force to converge quickly and synchronously to the target range, significantly shortening the time required for a single calibration and significantly improving the efficiency of maintenance operations.

[0016] This method features a highly automated and intelligent process, with data acquisition, feature recognition, deviation calculation, and adjustment execution all completed under the guidance and control of the system. This reduces reliance on the individual experience and skills of operators, enabling technicians with basic training to perform high-quality calibrations. It effectively reduces the risk of human error, achieving "de-skilling" and standardization of operations, which helps reduce training costs and ensures consistency of results across different personnel and times. Simultaneously, the entire process is digitally recorded, including complete force-position curves, adjustment history, and timestamps, providing tamper-proof data traceability for calibration work and meeting the stringent requirements of quality traceability and compliance auditing in the aviation maintenance field.

[0017] This method extracts multi-dimensional health characteristics from the force-position curve, such as smoothness, hysteresis, and segment decay, and combines this with longitudinal trend analysis of historical data and horizontal comparison with similar aircraft fleets. This enables the quantitative assessment of the real-time health status of the throttle lever force-sensing unit and the prediction of its performance degradation trajectory. This allows for a shift in maintenance from traditional "post-failure repair" or "routine maintenance" to condition-based "predictive maintenance," providing early warnings before potential failures occur and generating targeted maintenance recommendations. This not only prevents flight safety issues that may arise from abnormal throttle lever force sensing but also reduces unplanned downtime by optimizing maintenance timing, extending component lifespan and resulting in significant safety and economic benefits. Attached Figure Description

[0018] 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: Figure 1 This is a flowchart of a force-sensing calibration method for an aircraft throttle lever according to the first embodiment of the present invention; Figure 2 This is a schematic diagram of the connection of each module in a force-sensing calibration system for an aircraft throttle lever according to a second embodiment of the present invention. Detailed Implementation

[0019] The present application will now be described in further detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the invention. Furthermore, it should be noted that, for ease of description, only the parts relevant to the invention are shown in the accompanying drawings.

[0020] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0021] The first embodiment of the present invention provides a force-sensing calibration method for an aircraft throttle lever, the method comprising: Step S10: Synchronously acquire force and position signals of the throttle lever during its full stroke operation to generate a continuous force-position curve; Step S20: Based on the preset feature template, identify the abrupt change features of the positioning point in the force-position curve, and divide the entire stroke of the throttle lever into a positioning segment containing the abrupt change features of the positioning point and a smooth stroke segment located between the positioning points; Step S30: Calculate the average force signal value within the smooth travel segment as the basic friction force, extract the peak force signal value within the positioning segment as the positioning holding force, and calculate the deviation of the basic friction force and the positioning holding force from their respective preset target values. Step S40: Apply a preset amplitude test adjustment amount to the friction adjustment mechanism and the positioning adjustment mechanism of the throttle lever respectively, and measure the cross change amount of the basic friction force and the positioning holding force caused by the test action of the two adjustment mechanisms; Step S50: Construct an influence coefficient matrix using sensitivity coefficients; the sensitivity coefficients are calculated based on the ratio of the cross-variance to the experimental adjustment. Step S60: Calculate the inverse matrix of the influence coefficient matrix, perform operations on the deviation amount and the inverse matrix to calculate the target physical adjustment amount used to eliminate the deviation amount, and drive the friction adjustment mechanism and the positioning adjustment mechanism to operate simultaneously.

[0022] To more clearly illustrate the force-sensing calibration method for an aircraft throttle lever according to the present invention, the following is in conjunction with... Figure 1 The steps in the embodiments of the present invention are described in detail below: Step S10: Synchronously acquire force and position signals of the throttle lever during its full stroke operation to generate a continuous force-position curve; In this embodiment, the step of generating a continuous force-position curve specifically includes: Step S11: Simultaneously acquire force sensor data and position encoder data at a preset high sampling frequency, and perform timing alignment. Step S12: Apply an adaptive filtering algorithm to process the original force signal to filter out mechanical vibration noise and operational jitter noise; Step S13: Map the denoised force signal to the position coordinate axis to construct a continuous curve of force changing with position, and calculate the first derivative curve and the second derivative curve of the continuous curve with respect to position, wherein the first derivative curve is used to help identify the abrupt change rate of the force value.

[0023] In step S10, the process of synchronously acquiring force and position signals of the throttle lever during its full stroke to generate a continuous force-position curve is the data foundation of the entire calibration method. Its execution quality directly determines the accuracy of subsequent feature recognition and adjustment. Specifically, this step first involves high-precision data acquisition and timing alignment at the hardware level, i.e., step S11. At this stage, the operator, guided by the system, must push or pull the throttle lever at a preset, stable low speed (generally recommended to be 5 to 10 degrees per second) to complete at least two full-stroke operations, including the push and pull processes. At this time, the system activates the force sensing module and position sensing module to work synchronously. The system uses a non-invasive, high-precision force sensor to sense the operating resistance or holding force experienced by the throttle lever at various positions in real time. The sensor's accuracy is better than ±0.01 daN. Simultaneously, a high-resolution position encoder linked to the throttle mechanism with a resolution better than 0.1 degrees is used to accurately calibrate the current angular or linear displacement physical position of the throttle lever.

[0024] The data processing and control unit synchronously triggers data reading commands from the two sensors at a preset high sampling frequency of no less than 100Hz, ensuring that the current instantaneous force and position values ​​are captured simultaneously within an extremely short time window. To eliminate time misalignment caused by transmission delays or differences in system response, the system performs strict time-series alignment processing on the two acquired raw data streams, binding force and position data with the same timestamp one-to-one to form a raw high-density dataset containing force, position, and time information, thereby realistically reproducing the mechanical response trajectory of the gate arm during dynamic operation.

[0025] After acquiring the raw data, due to the unavoidable environmental interference and human factors in actual working conditions, step S12 needs to be executed, that is, the adaptive filtering algorithm is applied to process the raw force signal. The raw force sensor signal is often mixed with a variety of noise components, including but not limited to the background mechanical vibration noise of the aircraft when it is stopped, the electronic noise of the sensor circuit itself, and high-frequency spikes such as the operation jitter noise caused by the slight tremor of the operator's hand when manually pushing the throttle lever.

[0026] To filter out this noise without compromising the crucial details of the force signal that accurately reflect the mechanical characteristics, particularly the abrupt changes in force values ​​at the positioning points, this embodiment employs an adaptive Kalman filter algorithm. This algorithm adjusts the filter gain in real time based on the signal's statistical characteristics, effectively identifying and smoothing out high-frequency random interference caused by jitter and vibration. Simultaneously, it preserves the true force variation trend in the force-position curve and the rising or falling edge shape of the positioning point signal to the greatest extent possible, generating a smooth and high-fidelity clean force signal. This provides a data source with an extremely high signal-to-noise ratio for subsequent precise analysis.

[0027] In step S13, the denoised pure force signal is no longer considered a function of time, but rather mapped to the spatial domain, specifically onto the position coordinate axis, constructing a continuous force-position curve with the angle or displacement of position as the abscissa and the force value as the ordinate. Since the acquisition process includes travel in both the push and pull directions, the system constructs separate push and pull process curves for subsequent analysis of hysteresis characteristics. To enable the computer to automatically understand the characteristic meaning of the curves, the system further performs mathematical morphological processing on the continuous curve, calculating the first derivative curve of the continuous curve with respect to position using numerical difference or fitting differentiation methods. With the second derivative curve .

[0028] The first derivative curve characterizes the rate of change of force with position, i.e., the rate of change or slope of force, and is the core basis for subsequent identification of the positioning point, because the positioning point usually corresponds to a sharp increase or decrease in force in mechanical performance. The second derivative curve, on the other hand, characterizes the curvature or concavity of the force change, and is used in subsequent steps to evaluate the smoothness of the curve and the stability of mechanical operation. Through the above series of rigorous data acquisition, cleaning, mapping and transformation processes, the system completes the conversion from mechanical motion in the physical world to high-dimensional characteristic curves in the digital world, laying a solid data foundation for the intelligent calibration of throttle lever force.

[0029] Step S20: Based on the preset feature template, identify the abrupt change features of the positioning point in the force-position curve, and divide the entire stroke of the throttle lever into a positioning segment containing the abrupt change features of the positioning point and a smooth stroke segment located between the positioning points; In this embodiment, the step of dividing the full travel of the throttle lever into a positioning segment and a smooth travel segment specifically includes: Step S21: Establish a key point feature template that includes the range of force value peak amplitude, the span of wave width position, and the morphological characteristics of derivative. Step S22: Scan the first derivative curve, identify local extrema and zero intersections, and extract candidate mutation regions; Step S23: Perform normalized cross-correlation operation on the candidate mutation region and the feature template, calculate the morphological similarity, and confirm the effective localization point in combination with the travel position constraint; Step S24: Taking the confirmed valid positioning point as the center, extend a preset width to both sides to define the positioning segment, and define the continuous travel outside the positioning segment as the smooth travel segment.

[0030] Step S20, which involves identifying abrupt changes in the positioning point features of the force-position curve based on a preset feature template and dividing the stroke, is a crucial step in achieving refined and decoupled calibration of the throttle lever's force feel. The core of this step lies in using intelligent algorithms to deconstruct the complex full-stroke mechanical response of the throttle lever into different functional regions.

[0031] In practice, step S21 is executed first to establish a high-dimensional key point feature template library. This template library is not a simple set of force thresholds, but rather constructed based on the typical mechanical forms of various standard throttle lever positioning points pre-stored in the system (such as the Idle Stop position and the Max Climb position). Each feature template contains multiple dimensions of descriptive parameters: First, the force peak amplitude range defines the typical range of the breakthrough force at the positioning point. Second, the wave width position span describes the physical width occupied by the positioning mechanism on the stroke axis. More importantly, the derivative morphology characteristics are recorded. The system records the standard morphology of the positioning point on the first derivative of the force-position curve, typically exhibiting a double-pulse characteristic of first showing a positive peak (rapid increase in force) followed by a negative peak (rapid decrease in force), as well as curvature changes on the second derivative. Furthermore, the template includes hysteresis characteristic parameters to describe the difference patterns in the positioning point morphology during push-rod and pull-rod processes. This templated design gives the system an expert-like recognition capability, enabling it to distinguish between genuine mechanical positioning slots and occasional signal interference.

[0032] After establishing a complete template library, the system executes step S22, which calculates the first derivative curve of the continuous force-position curve generated in step S10. A full-stroke scan is performed. The first derivative reflects the rate of force change. The presence of a location point can cause a drastic change in force, resulting in significant local extrema on the derivative curve. Numerical analysis algorithms identify all local maxima, minima, and zero-crossing points (corresponding to the peaks of the original force curve) on the derivative curve, marking the neighborhoods of these feature points as candidate abrupt change regions. This process achieves preliminary screening of potential location points, enabling rapid identification of all possible force abrupt change locations, regardless of whether they are caused by the actual positioning mechanism or by operational vibrations or mechanical foreign objects.

[0033] Subsequently, to accurately remove false features and confirm valid localization points from the candidate regions, the system executes step S23, employing a composite recognition algorithm based on morphological feature matching. The system performs normalized cross-correlation calculations on the data fragments of each extracted candidate mutation region and a preset feature template. Through this calculation, the system can calculate the morphological similarity score between the candidate waveform and the standard template. The score, ranging from 0 to 1, quantifies the degree of shape matching between the two elements, unaffected by differences in absolute amplitude. Simultaneously, the system introduces stroke position constraints. As supplementary judgment criteria, for example, according to the aircraft design specifications, the MAX CLIMB positioning point must appear within a specific percentage range of the travel (e.g., 75%-85%). The system utilizes a comprehensive judgment function. The system performs a weighted calculation of morphological scores and positional constraints. Only when the comprehensive judgment value D exceeds a preset strict threshold (e.g., 0.85) will the candidate region be recognized as a valid localization point. In addition, the system has a self-learning function, which can dynamically update the template library with the features of each successfully identified localization point to adapt to individual differences between different aircraft.

[0034] Finally, in step S24, the entire stroke is physically divided according to the confirmed effective positioning point location. Using the geometric center location of the confirmed effective positioning point (i.e., the peak of the force value or the zero-crossing point of the derivative) as a reference, the system extends a preset physical width (set according to the mechanical design dimensions of the positioning mechanism) to both sides of the stroke, defining this area containing the entire process of force value abrupt changes as the positioning segment. Within the positioning segment, the change in force value is mainly dominated by the interaction between the positioning roller and the cam. Correspondingly, the system defines the continuous stroke outside the positioning segment, which eliminates all positioning abrupt change characteristics, as the smooth stroke segment. Within these smooth segments, the resistance experienced by the throttle lever mainly comes from the basic frictional force of the mechanical transmission chain. Through this precise segmentation, the system successfully achieves spatial separation of the basic frictional force and the positioning holding force, providing accurate data boundaries for subsequent steps to independently calculate deviations and decouple adjustments for these two parameters.

[0035] The following is a detailed example of how to confirm the "MAX CLIMB" positioning point through specific calculations. The system first calls the "MAX CLIMB" standard template from the preset feature template library. This template defines the standard position of the positioning point as being between 78% and 82% of the throttle lever travel and includes a standard normalized first-order derivative waveform consisting of 101 data points. This waveform typically exhibits a double-pulse pattern, first positive and then negative, accurately depicting the steep increase in force when breaking through the slot and the steep decrease when engaging the slot. When the system analyzes a candidate abrupt change region located at 79.5% of the travel, it collects the measured first-order derivative waveform data for that region and normalizes it to 101 points for comparison.

[0036] Morphological similarity is calculated using the Normalized Cross-Correlation (NCC) algorithm, and its formula is as follows: .in, This represents the i-th data point of the candidate waveform. In the example, it corresponds to the first derivative waveform data of the location point region obtained from actual measurement. This represents the i-th data point of the template waveform. The template waveform is a pre-stored reference waveform that represents the shape of a standard positioning point. The mean of candidate waveform C, i.e., all The average value. Used to center the waveform and eliminate the effect of overall amplitude offset. The mean of the template waveform T, i.e., all The average value. Also used for centering. n represents the total number of data points in the waveform.

[0037] In this example, assuming the sum of the mean-reduced dot products of the candidate waveform and the template waveform is 85, the standard deviation of the candidate waveform is 0.9, and the standard deviation of the template waveform is 1.0, substituting these values ​​into the formula yields the morphological similarity score. This high score indicates that the measured waveform closely matches the shape of the template.

[0038] Simultaneously, the position constraint score is calculated. The standard position interval center is defined as 80% according to the template, and the maximum allowable deviation is set to half the interval width, i.e., 2%. The actual position of the candidate point is 79.5%, with an absolute position deviation of 0.5%. A linear decay function is used. Perform the calculation and substitute the values. This indicates that the candidate point is located very close to the ideal center.

[0039] The system then weights and fuses these two scores for a comprehensive judgment, with morphological similarity having a specific weight. The position constraint weight is 0.7. The value is 0.3. Substituting this into the calculated score, that is... .

[0040] Finally, the system compares the calculated comprehensive judgment value of 0.886 with the preset strict threshold of 0.85. Since the judgment value clearly exceeds the threshold, the system confirms the candidate region as a valid "MAX CLIMB" location point. This complete calculation process demonstrates how to quantify multi-dimensional features such as waveform morphology and physical location into objective scores, and make a deterministic judgment through rigorous formula calculation and weighted fusion, thereby ensuring the robustness and repeatability of the recognition method.

[0041] Step S30: Calculate the average force signal value within the smooth travel segment as the basic friction force, extract the peak force signal value within the positioning segment as the positioning holding force, and calculate the deviation of the basic friction force and the positioning holding force from their respective preset target values. In step S30, the system will perform precise calculations and deviation analysis of the basic friction force and positioning holding force based on the high-fidelity force-position curve data processed by adaptive Kalman filtering in the previous steps. First, for the calculation of the basic friction force, the system needs to determine the effective "smooth travel segment", which is defined as the continuous travel portion of the throttle lever's full travel after removing abrupt changes in positioning points such as IDLE STOP and MAX (maximum climb position).

[0042] Specifically, based on the positioning point location identified in step S23, the data processing unit automatically defines the continuous travel segment between IDLE STOP and MAX CLIMB on the force-position curve as the analysis interval, and automatically cuts off the transition data of a preset width near these two endpoints to eliminate the interference of nonlinear force abrupt changes when the positioning mechanism enters and exits the slot on the average value calculation. Within this smooth segment, the system performs an arithmetic mean calculation on the force signal values ​​of all sampling points to calculate the measured average value of the basic friction force, denoted as . Simultaneously calculate the standard deviation of the internal force values ​​in this section. This is used to quantitatively assess the stability of the basic friction force. If the standard deviation exceeds the preset threshold, it indicates that there may be abnormal jamming or uneven lubrication in the mechanical structure.

[0043] Next, the preset model configuration parameters are called to obtain the first target force range of the basic friction force (e.g., 0.85 daN to 1.05 daN) and its central target value. By subtracting the measured average value from the central target value ( The deviation of the basic friction force is calculated, and the sign of this deviation indicates the required rotation direction of the subsequent adjusting nut of the first servo motor.

[0044] Meanwhile, the calculation of positioning holding force (i.e., breakthrough force) focuses on the identified positioning segment, especially the MAX CLIMB positioning point area. To overcome the random errors that may exist in a single measurement, a comprehensive analysis is performed based on 3 to 5 reciprocating over-position scanning data continuously executed by the operator at a moderate speed (approximately 15-20° / s).

[0045] In each scan of the positioning point, the peak force signal within the positioning segment is precisely extracted by finding the extreme value or local maximum value of the derivative. This peak value represents the instantaneous maximum force required for the throttle lever to break through the positioning slot. Multiple peak data points obtained from multiple scans were recorded. First, outlier detection was performed, removing abnormal measurements that deviated from the mean by more than three standard deviations. Then, the arithmetic mean of the remaining valid peak values ​​was calculated. As the final confirmed current positioning holding force, the standard deviation of these peak values ​​is calculated simultaneously. To verify the consistency of measurements (e.g., requirements) Subsequently, the system loads a second target force parameter (e.g., 2.5 daN, with an allowable tolerance of +0.1 / -0.0 daN), setting the target breakthrough force value. Subtract the calculated average breakthrough power value Thus, the deviation of the positioning and holding force is obtained. .

[0046] The calculated basic friction force deviation Deviation from positioning and holding force The deviation vector is constructed as a bivariate vector and used as the input parameter of the decoupling control algorithm. It is then used to accurately calculate the amount of coordinated adjustment between the two motors required to overcome the mechanical coupling effect.

[0047] Step S40: Apply a preset amplitude test adjustment amount to the friction adjustment mechanism and the positioning adjustment mechanism of the throttle lever respectively, and measure the cross change amount of the basic friction force and the positioning holding force caused by the test action of the two adjustment mechanisms; In this embodiment, the throttle lever system includes two core mechanical adjustment mechanisms: one is a friction adjustment mechanism, which is usually composed of a set of coaxially mounted friction plates, spacers and compression springs. The axial preload of the spring is changed by the adjusting nut located at the shaft end, thereby adjusting the normal pressure between the friction plates, and thus changing the basic friction force of the throttle lever throughout its stroke. Secondly, there is the positioning adjustment mechanism. This mechanism typically includes a roller or slider assembly that meshes with the throttle lever cam profile. By adjusting the screw, the compression of the positioning spring in this assembly or the engagement depth between the positioning roller and the cam groove can be changed, thereby altering the positioning holding force required for the throttle lever to break through a specific positioning point (such as MAX CLIMB). It should be noted that the specific mechanical structure and adjustment principle of the aforementioned friction adjustment mechanism and positioning adjustment mechanism are existing mature technologies in this field, and their internal structures will not be elaborated upon here.

[0048] In this embodiment, the first servo motor is first instructed to drive its dedicated sleeve tool head to precisely engage with the adjusting nut of the friction adjustment mechanism, and to apply a preset small rotation amplitude. For example, rotating 1 / 8 turn or a specific angle, this amplitude needs to be set within a range that causes an observable change in force without causing overshooting of the system state. After applying this action, the changes in two physical quantities are measured simultaneously: not only the change in the basic frictional force directly caused by the rotation of the adjusting nut. It also focuses on measuring the consequent changes in the positioning and holding force caused by unexpected movement of the adjusting nut due to rigid coupling or stress transmission in the mechanical structure. This process captures the unidirectional coupling effect of friction on positioning.

[0049] The second servo motor is instructed to drive its micro-stepping tool head to engage with the adjusting screw of the positioning adjustment mechanism, applying another preset micro-rotation amplitude. Similarly, the system performs synchronous measurement again, recording the change in positioning and holding force directly caused by the rotation of the adjusting screw. And the amount of reverse cross-effect on the base friction caused by the change in the preload of the positioning mechanism. Through the aforementioned step-by-step experimental adjustments and dual-channel measurements, the actual response characteristics of the two adjustment mechanisms under the current mechanical state and the cross-coupling data between them were accurately obtained. This provided the necessary physical measurement basis for the subsequent establishment of a mathematical model, thus solving the technical problem of repeated adjustments failing to converge in traditional methods due to the neglect of mechanical coupling.

[0050] Step S50: Construct an influence coefficient matrix using sensitivity coefficients; the sensitivity coefficients are calculated based on the ratio of the cross-variance to the experimental adjustment. The specific steps for constructing the influence coefficient matrix include: Step S51: Obtain the change value of basic friction force and the change value of positioning holding force caused by the friction adjustment mechanism when the test adjustment amount is applied, divide them by the test adjustment amount respectively to obtain the first set of sensitivity coefficients, which are used as the first column elements of the influence coefficient matrix; Step S52: Obtain the change value of basic friction force and the change value of positioning holding force caused by the application of test adjustment amount by the positioning adjustment mechanism, divide them by the test adjustment amount respectively to obtain the second set of sensitivity coefficients, which are used as the second column elements of the influence coefficient matrix; Step S53: Establish a two-dimensional influence coefficient matrix based on the elements in the first column and the elements in the second column.

[0051] In this embodiment, the specific steps for calculating the target physical adjustment amount include: Step S54: Calculate the pseudo-inverse matrix of the influence coefficient matrix, perform matrix multiplication on the pseudo-inverse matrix and the deviation vector composed of the basic friction force deviation and the positioning holding force deviation, and use the calculation result as the target physical adjustment vector containing the rotation adjustment amount of the friction adjustment mechanism and the rotation adjustment amount of the positioning adjustment mechanism.

[0052] After completing the micro-perturbation test in step S40 and obtaining the key cross-variable data, the method of this invention proceeds to step S50, which involves using the measured data from the previous stage to construct an influence coefficient matrix that can accurately describe the multivariable coupling characteristics of the system. The core of this step is to transform the physical mechanical response into a mathematical model that can be used by the control algorithm.

[0053] Specifically, firstly, based on the cross-variance recorded in step S40 and the known experimental adjustment, a set of sensitivity coefficients is calculated to quantify the relationship between the adjustment action and the force value change. The physical meaning of this sensitivity coefficient is the magnitude of the change in the target force value (basic friction force or positioning holding force) caused by a unit adjustment (e.g., rotating the adjusting nut or screw one revolution or a specific angle), which directly reflects the control effectiveness and coupling strength of the adjustment mechanism.

[0054] The construction process of the influence coefficient matrix is ​​specifically broken down into the following sequential sub-steps. First, step S51 is executed to retrieve the experimental adjustment amount applied to the friction adjustment mechanism. The measured change in basic friction force Cross-variance value of positioning and holding force Next, divide each of these two force changes by the applied test adjustment amount. Thus, the first set of sensitivity coefficients is calculated, i.e. and J11 characterizes the sensitivity of the friction adjustment mechanism to its primary control target—the basic friction force; while J21 quantifies the degree of cross-coupling influence of the friction adjustment mechanism on the positioning and holding force. These two calculated sensitivity coefficients, J11 and J21, together constitute the first column of the influence coefficient matrix J.

[0055] Subsequently, step S52 is executed seamlessly, and the system processes the test data of the positioning adjustment mechanism in exactly the same way. It retrieves the test adjustment amount applied to the positioning adjustment mechanism. The measured change in positioning holding force Cross-variance value of basic friction force The system divides each of the two force changes by the applied test adjustment. The second set of sensitivity coefficients was calculated, namely and Among them, J22 characterizes the direct sensitivity of the positioning adjustment mechanism to its main control target—the positioning holding force; while J12 quantifies the reverse coupling effect of the positioning adjustment mechanism on the base friction force during adjustment. These two sensitivity coefficients, J12 and J22, constitute the influence coefficient matrix. The second column element.

[0056] Finally, in step S53, the system establishes a complete two-dimensional influence coefficient matrix based on the quaternary sensitivity coefficients calculated in the previous two sub-steps. The matrix is ​​in the following form: This matrix comprehensively and quantitatively describes the two inputs (the adjustment amount of the friction regulating mechanism). Adjustment amount of the positioning adjustment mechanism The linear relationship between the throttle lever and its two outputs (the change in basic friction force and the change in holding force) is established, where the main diagonal elements represent direct control actions, and the off-diagonal elements represent coupled interference actions that need to be precisely compensated for. This results in a highly customized dynamic characteristic mathematical model tailored to the specific mechanical state of the throttle lever.

[0057] After successfully constructing the influence coefficient matrix Next, the method of the present invention further executes step S54, using the matrix model to accurately calculate the physical adjustment required to achieve the target force value. First, it is necessary to determine the deviation between the current throttle lever force feel and the target standard. By collecting the current average value of the basic friction force and the average value of the positioning holding force, and comparing them with the preset target value (or the center value of the target range), the basic friction force deviation is obtained. and positioning holding force deviation These two deviations constitute a two-dimensional deviation vector. This vector indicates the "distance" and "direction" between the current state and the target state.

[0058] Next, in order to determine the adjustment amount required to eliminate the deviation vector E in one step, the pseudo-inverse matrix of the newly established influence coefficient matrix J is calculated, resulting in... Using a pseudo-inverse matrix instead of a standard inverse matrix enhances the algorithm's robustness when dealing with singular or ill-conditioned matrices. Obtaining the pseudo-inverse matrix... Then, the system performs the core decoupling calculation: the pseudo-inverse matrix is... Perform matrix multiplication with the deviation vector E, i.e. The result of the calculation is a two-dimensional target physics adjustment vector. = Due to the pseudo-inverse matrix The calculation process already includes decoupling of the coupling effect; therefore, the adjustment amount obtained through this single matrix operation is... and It is feedforward compensated, which means that performing these two adjustment actions can theoretically and simultaneously and accurately calibrate the basic friction force and the positioning holding force to the target value, thereby avoiding repeated iterations and over-adjustment caused by coupling effect in traditional adjustment methods, and greatly improving the efficiency and accuracy of calibration.

[0059] Step S60: Calculate the inverse matrix of the influence coefficient matrix, perform operations on the deviation amount and the inverse matrix to calculate the target physical adjustment amount used to eliminate the deviation amount, and drive the friction adjustment mechanism and the positioning adjustment mechanism to operate simultaneously.

[0060] The specific steps for simultaneously driving the friction adjustment mechanism and the positioning adjustment mechanism include: Step S61: Using a variable gain control strategy, the calculated target physical adjustment amount is converted into a control amount for the servo motor according to different proportions based on the current force deviation amount. Step S62: When the absolute value of the deviation is greater than the first preset threshold, a large adjustment is performed using the first gain coefficient. Step S63: When the absolute value of the deviation is less than or equal to the first preset threshold but greater than the second preset threshold, a second gain coefficient less than the first gain coefficient is used to perform fine adjustment; Step S64: After one or more adjustment operations, re-acquire the force-position curve and calculate the current force value. If the basic friction force or positioning holding force deviates from the preset tolerance range due to the fixed operation after adjustment, a new round of calculation and adjustment will be automatically started based on the new deviation until the two force value parameters are simultaneously stable within the preset tolerance range.

[0061] In step S61, the drive control of this invention employs a variable gain control strategy. The core idea of ​​this strategy is to dynamically adjust the actual applied physical adjustment amount based on the absolute value of the current force deviation, i.e., the target physical adjustment amount calculated in the previous step. The final control output is converted into servo motors at different ratios.

[0062] Specifically, the variable gain strategy includes the following discrimination and execution logic. First, in step S62, a first preset threshold representing a "large deviation" state is set (e.g., deviation exceeding 50% of the target tolerance range). When any force deviation ( or When the absolute value of the first gain coefficient is greater than this first preset threshold, the system determines that it is currently in the "coarse adjustment" stage, which is far from the target value. At this time, the system uses a larger first gain coefficient (e.g., (Between 0.8 and 1.0), the calculated target physical adjustment is multiplied by this coefficient and then sent to the motor for execution. This operation aims to utilize a larger adjustment step size to enable the system state to converge quickly to the target region, thereby significantly shortening the time spent in the early stages of calibration.

[0063] In step S63, a second preset threshold representing the "fine-tuning zone" is also set (e.g., the deviation is 10% of the target tolerance range), which is much smaller than the first preset threshold. When the absolute value of the force deviation has decreased to less than or equal to the first preset threshold, but is still greater than the second preset threshold, the system determines that it has entered the "fine-tuning" stage requiring fine-tuning. In this stage, the system automatically switches to a second gain coefficient much smaller than the first gain coefficient (e.g., ...). (Between 0.3 and 0.5). Using a smaller gain coefficient means smaller adjustment steps, which helps suppress system inertia and prevents overshooting of the target range, thus achieving a smooth and accurate approximation of the target value.

[0064] The entire automatic calibration process is a rigorous closed-loop iterative cycle, as described in step S64. After performing one or more coordinated adjustment operations based on a variable gain strategy, the operator is immediately prompted to perform a full-stroke scan again to acquire the adjusted new force-position curve. The data processing and control unit recalculates the current basic friction force and positioning holding force and evaluates whether they meet the standards. In particular, this invention fully considers that after adjustment, the operator needs to manually tighten the locking mechanism of the adjusting nut according to the procedure. The stress generated by this tightening action may cause a slight deviation in the calibrated force value. Therefore, after final tightening, the system will also force a verification scan. If the basic friction force or positioning holding force is found to deviate from the preset tolerance range for any reason (including the tightening operation), the system will treat it as a new initial state and automatically start a new round of "calculation-adjustment-verification" cycle based on this new deviation. This cycle will continue until the system confirms that both force parameters can be stably maintained within their respective preset tolerance ranges in the final locked state. This adaptive iteration and verification mechanism, combined with the aforementioned decoupled feedforward and variable gain feedback control, ensures that the present invention not only has an efficient adjustment process, but also accurate and reliable final results, and can completely overcome the various uncertainties present in traditional manual calibration.

[0065] In this embodiment, the method further includes: Extract health characteristic parameters from the calibrated force-position curve; By combining the health characteristic parameters of this calibration with historical calibration data, a time series analysis was conducted to calculate the health index, which characterizes the degree of performance degradation of the throttle lever force sensor unit. Based on the current value and trend of the health index, predictive maintenance recommendations are generated.

[0066] Specifically, the steps for extracting the health characteristic parameters of the calibrated force-position curve include: The dispersion index of the second derivative of the force-position curve within the smooth stroke section is calculated and used as a smoothness index characterizing the mechanical smoothness. Obtain the push rod process curve and the pull rod process curve respectively, and calculate the integral area of ​​the force difference between the two at the same stroke position over the stroke, which is used as the hysteresis loop parameter characterizing mechanical clearance and internal friction. The stroke is divided into several sub-intervals, and the attenuation ratio of the average force value of each sub-interval relative to the historical initial state of the throttle lever is calculated as the segment attenuation rate parameter. The peak value of the force change rate of the positioning holding force on the rising edge is calculated as a steepness parameter characterizing the roller wear state of the positioning mechanism.

[0067] The specific steps for calculating the health index, which characterizes the degree of performance degradation of the throttle lever force sensor unit, include: The characteristic parameters in the historical calibration records of the aircraft were retrieved, and the time series analysis method was applied to fit the parameter evolution trend line to obtain the slope of the trend line. Retrieve the statistical distribution of characteristic parameters of the same aircraft type group, and calculate the deviation score of the current characteristic parameter relative to the group mean. Using the standard reference curve provided by the manufacturer as a reference, calculate the dynamic time bending distance between the current force-position curve and the reference curve; The slope of the trend line, the deviation score, and the dynamic time bending distance are weighted and fused to generate a normalized health index.

[0068] Based on the current value and trend of the health index, the specific steps for generating predictive maintenance recommendations include: The extracted health feature parameters are compared with preset fault feature thresholds or historical benchmark ranges. When the smoothness index exceeds the first abnormal threshold, maintenance suggestions are generated to check the lubrication status of mechanical moving parts or the presence of foreign objects. When the hysteresis loop area parameter exceeds the second abnormal threshold, maintenance recommendations for checking transmission mechanism clearance or bearing wear are generated. When the force attenuation rate of a specific stroke sub-segment exceeds the third abnormal threshold, maintenance recommendations are generated to check the wear status of the friction pads in the corresponding area. Based on the predicted time when the health index will drop to the preset warning threshold, a preventative replacement or in-depth maintenance work order is generated.

[0069] In a preferred embodiment of the present invention, after the basic friction force and positioning holding force of the throttle lever are successfully calibrated to the preset tolerance range through a decoupling control strategy, the system does not terminate immediately. Instead, it automatically utilizes the full-stroke, high-sampling-rate force-position curve data recorded in real time during the calibration process to enter a deep health assessment and predictive maintenance phase. This phase is designed to upgrade the maintenance mode from traditional responsive repair to proactive predictive maintenance, quantifying the degree of intrinsic performance degradation of the throttle lever force-sensing unit by exploring the subtle mechanical characteristics contained in the force-position curve.

[0070] First, multi-dimensional health feature parameters are extracted, which reflect the physical state of the mechanical structure from different perspectives. Specifically, for the identified smooth travel segments, the system calculates the second derivative of the force-position curve with respect to position. Since the smooth section reflects a relatively stable friction process, the dispersion of the second derivative (usually characterized by variance or standard deviation) can sensitively detect irregular jitters or minor jamming in mechanical motion. This index is defined as a smoothness index; if the second derivative fluctuates drastically, it indicates that the mechanical transmission surface may have lubricant drying, minor foreign matter, or uneven wear. Simultaneously, the system extracts and aligns the push rod process curve and pull rod process curve in a complete cycle, and calculates the area of ​​the closed loop formed by these two curves on the spatial coordinate axis. Physically, this difference in force value due to the change in motion direction essentially reflects the cumulative clearance and internal static friction of the mechanical transmission chain. This area is defined as a hysteresis loop parameter; its continuous increase is a direct indication of bearing housing loosening, connecting rod joint wear, or increased mechanical clearance.

[0071] In addition to global features, the system also performs local fingerprint feature analysis. The system pre-divides the throttle lever's full travel into several equally wide sub-intervals, for example, dividing it into 10-degree increments, and calculates the average force value within each sub-interval after calibration. These measured averages are then compared one by one with the initial delivery state or first-inspection data after major overhaul stored in the device's non-volatile memory to calculate the force attenuation ratio, thereby generating segment attenuation rate parameters. This set of parameters can accurately pinpoint the specific location of mechanical wear. For example, if analysis reveals significant force attenuation only in the low throttle range, it can be inferred that the corresponding friction pads in that area are more severely damaged. Furthermore, for positioning points such as MAX CLIMB, the system focuses on the force change rate along its rising edge, extracting the peak value of dF / dθ as a steepness parameter. A healthy positioning mechanism should have a steep force jump, and as the contact point between the positioning roller and the cam profile wears, the slope of this rising edge gradually becomes gentler.

[0072] After completing the feature extraction described above, the system proceeds to the fusion calculation of the Health Index (HI), a comprehensive evaluation process integrating longitudinal time dimensions, lateral fleet dimensions, and absolute benchmark dimensions. First, it retrieves all historical calibration feature records for the specific aircraft over a given period, uses an ARIMA model or exponential smoothing algorithm to fit the evolution trend line of the parameters, and extracts the slope to assess the degradation rate. Second, the system obtains the statistical distribution of feature parameters of the same aircraft type from the central server via a wireless data link, calculates the deviation score of each parameter relative to the fleet average (i.e., Z-score), and thus identifies the relative performance ranking of the throttle lever within the entire fleet. To quantify overall morphological deviations, the system introduces the Dynamic Time Warping (DTW) algorithm to calculate the minimum alignment distance between the current force-position curve and the standard benchmark curve provided by the manufacturer. Compared to simple Euclidean distance, the DTW algorithm can better compensate for small phase deviations in mechanical positioning points. Finally, the system performs a weighted fusion calculation of the trend line slope, deviation score, and DTW distance to generate a normalized health index HI. The score is usually set between 0 and 100, which intuitively shows the overall health level of the force unit.

[0073] Finally, based on the current HI value and its predicted downward trend, targeted predictive maintenance recommendations are automatically generated. This process is achieved by comparing the extracted health characteristic parameters with preset multi-level alarm thresholds or historical benchmark ranges in real time. When the smoothness index exceeds the first abnormal threshold, the system indicates that there may be local dry friction or fine debris, and recommends performing lubrication maintenance or cleaning checks. When the hysteresis loop area parameter exceeds the second abnormal threshold and the trend is upward, the system determines that the transmission clearance is excessive and recommends checking the bearing wear or the tightness of the transmission pins. When the force attenuation rate of a specific section reaches the warning value, the system generates recommendations to check the wear of the friction plates in that specific area. The core predictive function is that the system predicts the estimated remaining time for the HI value to drop to the critical warning threshold based on a trend fitting model. For example, if it is predicted that the HI will drop below 60 points within the next 300 flight cycles, a preventive replacement or in-depth maintenance work order will be generated in advance in the Maintenance Management Information (MMIS). This closed-loop system, encompassing data acquisition and intelligent decision-making, not only ensures long-term stability of the throttle lever's force feel but also significantly reduces the risk of unplanned aircraft downtime by optimizing maintenance timing.

[0074] To more specifically illustrate the health assessment and predictive maintenance functions of this invention, a concrete virtual example will be used below. Assume an Airbus A320 aircraft with registration number B-1234 has just completed its automated throttle lever force-sensing calibration. After successful calibration, it automatically enters the health assessment program and performs in-depth analysis of the newly acquired force-position curve. The system first extracts four key health characteristic parameters: calculating the second derivative variance of the smooth travel segment, i.e., the smoothness index, is 0.08 N. 2 / deg 2 This is significantly higher than the baseline value of 0.03 N recorded at the time of the aircraft's initial delivery. 2 / deg 2 By integrating the area of ​​the hysteresis loop formed by the push-pull process curves, the hysteresis loop parameter was found to be 1.5 N·deg, which is significantly larger than 1.1 N·deg six months ago. The system divides the stroke into nine 10-degree sub-intervals and finds that the average force attenuation rate in the high thrust setting interval of 70°-80° reaches -8.5%, while the attenuation rate in other intervals is within -3%. Finally, the system calculates the force rise steepness parameter of the MAX CLIMB positioning point to be 1.2 N / deg, which is significantly lower than the 1.8 N / deg in its "golden curve".

[0075] Subsequently, multi-dimensional information fusion was used to calculate the Comprehensive Health Index (HI). In the longitudinal time series analysis, the system retrieved calibration data from 10 times over the past two years for the B-1234 aircraft, finding that its HI showed a stable linear trend of decreasing by approximately 0.5 percentage points per 100 flight cycles, while the hysteresis loop area increased in an almost exponential manner. In the horizontal fleet comparison, the system queried data from over 200 A320 aircraft of the same age in the fleet database, calculating that the hysteresis loop area parameter of B-1234 was located at the 85th percentile of the fleet, with a Z-score as high as +1.8, indicating that its internal friction and clearance issues were already at a relatively high level among similar aircraft. In the comparison with the absolute benchmark, the Dynamic Time Warping (DTW) algorithm was used to calculate that the morphological distance between the current force-position curve and the "golden curve" at the time of manufacture was 45.2, far exceeding the normal wear range of 20. Finally, the system uses a preset weighted model to integrate the analysis results of the above dimensions, including the trend line slope, Z-score, DTW distance, and the normalized scores of each feature parameter, and finally calculates the current health index HI of the throttle lever force sensing unit to be 68 points.

[0076] Based on the HI score of 68 and its downward trend, the health status of the throttle lever was rated as "Yellow - Recommended Attention," triggering the generation logic for predictive maintenance recommendations. The system first performed root cause inference: the significantly increased hysteresis loop area, which was higher than average in the fleet, combined with the deterioration of the smoothness index, strongly pointed to poor bearing lubrication or wear in the transmission mechanism; simultaneously, the decrease in the positioning point steepness parameter clearly corresponded to wear on the contact surface between the positioning roller and the cam. The regression prediction model, based on the historical rate of HI decline, predicted that the index would fall below the 60-point threshold representing "Orange - Recommended Preventive Maintenance" after approximately 450 flight cycles. Based on the above information, the system automatically generated a specific, forward-looking electronic work order in the airline's Maintenance Management Information System (MMIS), as follows: "

Predictive Maintenance Work Order

[0077] Through this example, the present invention transforms abstract data into concrete and actionable maintenance decisions, achieving a leap from passive maintenance to proactive prediction, and greatly improving maintenance efficiency and flight safety.

[0078] See Figure 2 A force-sensing calibration system for an aircraft throttle lever according to a second embodiment of the present invention is used to perform the force-sensing calibration method for an aircraft throttle lever described above. The system includes: The data acquisition module is configured to simultaneously acquire force and position signals of the throttle lever during its full stroke operation, generating a continuous force-position curve. The interval division module is configured to identify the abrupt change features of the positioning points in the force-position curve based on a preset feature template, and divide the entire stroke of the throttle lever into a positioning segment containing the abrupt change features of the positioning points and a smooth stroke segment located between the positioning points. The deviation calculation module is configured to calculate the average force signal within the smooth stroke section as the basic friction force, extract the peak force signal within the positioning section as the positioning holding force, and calculate the deviation of the basic friction force and the positioning holding force from their respective preset target values. The change calculation module is configured to apply a preset amplitude test adjustment to the friction adjustment mechanism and the positioning adjustment mechanism of the throttle lever respectively, and measure the cross change in the basic friction force and the positioning holding force caused by the test actions of the two adjustment mechanisms; A matrix construction module is configured to construct an influence coefficient matrix using sensitivity coefficients; the sensitivity coefficients are calculated based on the ratio of the cross-variance to the experimental adjustment. The control module is configured to calculate the inverse matrix of the influence coefficient matrix, perform operations on the deviation amount and the inverse matrix to calculate the target physical adjustment amount used to eliminate the deviation amount, and drive the friction adjustment mechanism and the positioning adjustment mechanism to operate simultaneously.

[0079] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working process and related explanations of the methods described above can be found in the corresponding processes in the foregoing system embodiments, and will not be repeated here.

[0080] It should be noted that the force-sensing calibration system for an aircraft throttle lever provided in the above embodiments is only an example illustrating the division of the functional modules. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the modules or steps in the embodiments of the present invention can be further decomposed or combined. For example, the modules in the above embodiments can be merged into one module, or further divided into multiple sub-modules to complete all or part of the functions described above. The names of the modules and steps involved in the embodiments of the present invention are only for distinguishing the various modules or steps and are not considered as an improper limitation of the present invention.

[0081] A device according to a third embodiment of the present invention includes: At least one processor; and a memory communicatively connected to at least one of the processors; The memory stores instructions that can be executed by the processor to implement the force-sensing calibration method for an aircraft throttle lever described above.

[0082] A fourth embodiment of the present invention provides a computer-readable storage medium storing computer instructions, which are executed by the computer to implement the above-described force-sensing calibration method for an aircraft throttle lever.

[0083] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working process and related descriptions of the storage device and processing device described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0084] The terms “first”, “second”, etc., are used to distinguish similar objects, not to describe or indicate a specific order or sequence.

[0085] The term "comprising" or any other similar term is intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus / device that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent in such process, method, article, or apparatus / device.

[0086] The technical solution of the present invention has been described above with reference to the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will all fall within the scope of protection of the present invention.

Claims

1. A force-sensing calibration method for an aircraft throttle lever, characterized in that, The method includes: Simultaneously acquire force and position signals of the throttle lever during its full stroke operation to generate a continuous force-position curve; Based on the preset feature template, the positioning point abrupt change features in the force-position curve are identified, and the entire stroke of the throttle lever is divided into positioning segments containing positioning point abrupt change features and smooth stroke segments located between positioning points. The average force signal within the smooth travel segment is calculated as the basic friction force, and the peak force signal within the positioning segment is extracted as the positioning holding force. The deviations of the basic friction force and the positioning holding force from their respective preset target values ​​are calculated. A preset amplitude test adjustment amount is applied to the friction adjustment mechanism and the positioning adjustment mechanism of the throttle lever respectively, and the cross change in the basic friction force and the positioning holding force caused by the test action of the two adjustment mechanisms is measured. An influence coefficient matrix is ​​constructed using sensitivity coefficients; the sensitivity coefficients are calculated based on the ratio of the cross-variance to the experimental adjustment. The inverse matrix of the influence coefficient matrix is ​​calculated, and the deviation is operated on with the inverse matrix to solve for the target physical adjustment amount used to eliminate the deviation, and the friction adjustment mechanism and the positioning adjustment mechanism are driven to operate simultaneously.

2. The force-sensing calibration method for an aircraft throttle lever according to claim 1, characterized in that, The specific steps for generating a continuous force-position curve include: Force sensor data and position encoder data are acquired synchronously at a preset high sampling frequency and time-aligned. An adaptive filtering algorithm is applied to process the original force signal to filter out mechanical vibration noise and operational jitter noise; The denoised force signal is mapped to the position coordinate axis to construct a continuous curve of force changing with position. The first and second derivative curves of the continuous curve with respect to position are calculated, where the first derivative curve is used to help identify the abrupt change rate of the force value.

3. The force-sensing calibration method for an aircraft throttle lever according to claim 2, characterized in that, The specific steps for dividing the full travel of the throttle lever into a positioning segment and a smooth travel segment include: Establish a key feature template that includes the range of force value peak amplitude, the span of wave width position, and the shape characteristics of derivative; Scan the first derivative curve to identify local extrema and zero intersections, and extract candidate mutation regions; The candidate mutation regions and feature templates are cross-correlated and normalized to calculate morphological similarity. Valid localization points are then confirmed by combining the travel position constraints. The positioning segment is defined as the position that is confirmed as a valid positioning point, and the continuous travel outside the positioning segment is defined as the smooth travel segment.

4. The force-sensing calibration method for an aircraft throttle lever according to claim 1, characterized in that, The specific steps for constructing the influence coefficient matrix include: The changes in basic friction force and positioning holding force caused by the application of the test adjustment amount of the friction adjustment mechanism are obtained, and then divided by the test adjustment amount to obtain the first set of sensitivity coefficients, which are used as the first column elements of the influence coefficient matrix. The changes in basic friction force and positioning holding force caused by the application of the test adjustment amount of the positioning adjustment mechanism are obtained, and then divided by the test adjustment amount to obtain the second set of sensitivity coefficients, which are used as the second column elements of the influence coefficient matrix. A two-dimensional influence coefficient matrix is ​​established based on the elements in the first column and the elements in the second column.

5. The force-sensing calibration method for an aircraft throttle lever according to claim 1, characterized in that, The specific steps for calculating the target physical adjustment amount include: Calculate the pseudo-inverse matrix of the influence coefficient matrix, perform matrix multiplication on the pseudo-inverse matrix and the deviation vector composed of the basic friction force deviation and the positioning holding force deviation, and use the calculation result as the target physical adjustment vector containing the rotation adjustment amount of the friction adjustment mechanism and the rotation adjustment amount of the positioning adjustment mechanism.

6. The force-sensing calibration method for an aircraft throttle lever according to claim 1, characterized in that, The specific steps for simultaneously driving the friction adjustment mechanism and the positioning adjustment mechanism include: A variable gain control strategy is adopted, which converts the calculated target physical adjustment amount into the control amount of the servo motor according to different proportions based on the current force deviation amount; When the absolute value of the deviation is greater than the first preset threshold, a large adjustment is performed using the first gain coefficient. When the absolute value of the deviation is less than or equal to the first preset threshold but greater than the second preset threshold, a second gain coefficient less than the first gain coefficient is used to perform fine adjustment. After one or more adjustment operations, the force-position curve is re-acquired and the current force value is calculated. If the basic friction force or positioning holding force deviates from the preset tolerance range due to the fixed operation after adjustment, a new round of calculation and adjustment is automatically started based on the new deviation until the two force parameters are simultaneously stable within the preset tolerance range.

7. The force-sensing calibration method for an aircraft throttle lever according to claim 1, characterized in that, This method also includes: Extract health characteristic parameters from the calibrated force-position curve; By combining the health characteristic parameters of this calibration with historical calibration data, a time series analysis was conducted to calculate the health index, which characterizes the degree of performance degradation of the throttle lever force sensor unit. Based on the current value and trend of the health index, predictive maintenance recommendations are generated.

8. The force-sensing calibration method for an aircraft throttle lever according to claim 7, characterized in that, The specific steps for extracting the health characteristic parameters of the calibrated force-position curve include: The dispersion index of the second derivative of the internal force-position curve of the smooth stroke section is calculated and used as a smoothness index to characterize the mechanical smoothness. Obtain the push rod process curve and the pull rod process curve respectively, and calculate the integral area of ​​the force difference between the two at the same stroke position over the stroke, which is used as the hysteresis loop parameter characterizing mechanical clearance and internal friction; The stroke is divided into several sub-intervals, and the attenuation ratio of the average force value of each sub-interval relative to the historical initial state of the throttle lever is calculated as the segment attenuation rate parameter. The peak value of the force change rate of the positioning holding force on the rising edge is calculated as a steepness parameter characterizing the roller wear state of the positioning mechanism.

9. A force-sensing calibration method for an aircraft throttle lever according to claim 8, characterized in that, The specific steps for calculating the health index, which characterizes the degree of performance degradation of the throttle lever force sensor unit, include: The characteristic parameters in the historical calibration records of the aircraft were retrieved, and the time series analysis method was applied to fit the parameter evolution trend line to obtain the slope of the trend line. Retrieve the statistical distribution of characteristic parameters of the same aircraft type group, and calculate the deviation score of the current characteristic parameter relative to the group mean. Using the standard reference curve provided by the manufacturer as a reference, calculate the dynamic time bending distance between the current force-position curve and the reference curve; The slope of the trend line, the deviation score, and the dynamic time bending distance are weighted and fused to generate a normalized health index.

10. A force-sensing calibration method for an aircraft throttle lever according to claim 9, characterized in that, Based on the current value and trend of the health index, the specific steps for generating predictive maintenance recommendations include: The extracted health feature parameters are compared with preset fault feature thresholds or historical benchmark ranges. When the smoothness index exceeds the first abnormal threshold, maintenance suggestions are generated to check the lubrication status of mechanical moving parts or the presence of foreign objects. When the hysteresis loop area parameter exceeds the second abnormal threshold, maintenance recommendations for checking transmission mechanism clearance or bearing wear are generated. When the force attenuation rate of a specific stroke sub-segment exceeds the third abnormal threshold, maintenance recommendations are generated to check the wear status of the friction pads in the corresponding area. Based on the predicted time when the health index will drop to the preset warning threshold, a preventative replacement or in-depth maintenance work order is generated.

Images