Self-adaptive wind-resistant slope roof photovoltaic tile system and transmission mechanism fault diagnosis method thereof

Abstract

The invention provides a self-adaptive wind-resistant slope roof photovoltaic tile system and a transmission mechanism fault diagnosis method thereof, and the method comprises the steps: initializing a digital twinborn body of a physical transmission mechanism, and carrying out the multi-body dynamics prediction of a control instruction sequence of a drive device based on the digital twinborn body; a simulation angle response sequence and a simulation vibration response sequence of the physical transmission mechanism in the current anti-wind adjustment action are obtained, and then the transmission ratio offset coefficient and the structure resonance abnormal index of the physical transmission mechanism are determined; determining response matching characteristics and excitation conflict characteristics of the physical transmission mechanism according to the transmission ratio offset coefficient and the structural resonance abnormity index; and carrying out abnormity judgment on the physical transmission mechanism based on the response matching characteristics, and positioning the fault position of the physical transmission mechanism through the excitation conflict characteristics when abnormity is judged to exist. Based on the scheme, fault diagnosis based on digital twinning virtual-real comparison can be realized.

Description

Technical Field

[0001] This application relates to the field of fault detection technology, and more specifically, to an adaptive wind-resistant sloped roof photovoltaic tile system and a fault diagnosis method for its transmission mechanism. Background Technology

[0002] Roof photovoltaic tiles are an integrated building material product that combines photovoltaic power generation with traditional building roof tiles. They adopt a tile-style design that complies with building codes and can directly replace traditional roof tiles. While achieving basic building functions such as wind and rain protection and heat insulation, they convert solar energy into electrical energy through built-in high-efficiency solar cell modules.

[0003] Current technologies for diagnosing transmission mechanism faults employ single-dimensional signal analysis methods. Whether relying on spectral analysis of vibration signals to identify impact components or monitoring transmission errors through angle encoder signals, these are essentially isolated observations of a specific physical phenomenon. This isolated analysis mode reveals significant limitations when facing early, minor, and complex potential faults. A single signal dimension cannot fully characterize the synergistic degradation effects caused by such faults across the two coupled physical dimensions of kinematic transmission accuracy and structural dynamics. Minor wear may not yet produce obvious sidebands in the vibration spectrum, but it can already cause slight fluctuations in the transmission ratio; while a slight increase in clearance may be masked at conventional angular resolution but exhibit phase lag in transient impact responses. This makes single-signal-based analysis methods prone to missing or misjudging the evolution trend of early complex faults, failing to provide sufficiently sensitive and reliable evidence for predictive maintenance. Therefore, how to achieve fault diagnosis based on digital twin virtual-real comparison, thereby improving the accuracy of transmission mechanism fault detection and location, has become a challenge for the industry. Summary of the Invention

[0004] This application provides an adaptive wind-resistant pitched roof photovoltaic tile system and a fault diagnosis method for its transmission mechanism, which can realize fault diagnosis based on digital twin virtual-real comparison, thereby improving the accuracy of fault detection and location of the transmission mechanism.

[0005] In a first aspect, this application provides a method for diagnosing transmission mechanism faults in an adaptive wind-resistant roof photovoltaic tile system, comprising: When the physical transmission mechanism performs adaptive wind resistance adjustment, the control command sequence of the drive device, the actual angle response sequence of the driven device, and the actual vibration response sequence are collected. Initialize the digital twin of the physical transmission mechanism, and perform multibody dynamics prediction on the control command sequence based on the digital twin to obtain the simulated angle response sequence and simulated vibration response sequence of the physical transmission mechanism in the current wind resistance adjustment action; The simulated angle response sequence and the actual angle response sequence are compared in time and frequency to obtain the transmission ratio offset coefficient of the physical transmission mechanism. At the same time, the simulated vibration response sequence and the actual vibration response sequence are evaluated for frequency band difference to obtain the structural resonance anomaly index of the physical transmission mechanism. Then, the response matching characteristics and excitation conflict characteristics of the physical transmission mechanism are determined based on the transmission ratio offset coefficient and the structural resonance anomaly index. Based on the response matching features, anomalies are determined in the physical transmission mechanism. When an anomaly is determined, the fault location of the physical transmission mechanism is located through the excitation conflict features.

[0006] In some embodiments, multibody dynamics prediction of the control command sequence based on the digital twin to obtain the simulated angle response sequence and simulated vibration response sequence of the physical transmission mechanism in the current wind-resistant adjustment action specifically includes: The control command sequence is converted into a time series of the equivalent input torque and speed of the drive unit; In the digital twin, based on the geometric parameters, inertial parameters, stiffness parameters and damping parameters of the transmission chain, a multi-degree-of-freedom dynamic equation is established, which includes gear pair meshing, bearing support and the flexibility of connecting parts. Using the equivalent input torque and rotational speed in the time series as boundary conditions, the dynamic equation is solved numerically with time steps, and the simulated angle response sequence and simulated vibration response sequence are output synchronously.

[0007] In some embodiments, performing a time-frequency joint comparison between the simulated angle response sequence and the actual angle response sequence to obtain the transmission ratio offset coefficient of the physical transmission mechanism specifically includes: The simulated angle response sequence and the actual angle response sequence are synchronously resampled at equal intervals, and the simulated value and the actual value of the instantaneous transmission ratio are calculated respectively. In the time domain, the absolute difference between the actual and simulated values ​​of the instantaneous transmission ratio at each sampling moment is calculated to obtain the transmission ratio deviation sequence; In the frequency domain, a fast Fourier transform is performed on the transmission ratio deviation sequence to extract the peak value of the amplitude spectrum at the transmission characteristic frequency. The transmission ratio offset coefficient of the physical transmission mechanism is determined by weighting the root mean square value of the transmission ratio deviation sequence and the peak value of the amplitude spectrum at the characteristic frequency.

[0008] In some embodiments, evaluating the frequency band difference between the simulated vibration response sequence and the actual vibration response sequence to obtain the structural resonance anomaly index of the physical transmission mechanism specifically includes: Wavelet packet decomposition is performed on both the simulated vibration response sequence and the actual vibration response sequence, decomposing them to a preset frequency band level. Calculate the energy ratio of the actual vibration signal to the simulated vibration signal in each corresponding frequency band of the frequency band hierarchy, and then obtain the energy logarithmic ratio of each frequency band in the frequency band hierarchy; The structural resonance anomaly index of the physical transmission mechanism is determined by all energy logarithmic ratios.

[0009] In some embodiments, determining the response matching characteristics and excitation conflict characteristics of the physical transmission mechanism based on the transmission ratio offset coefficient and the structural resonance anomaly index specifically includes: The transmission ratio offset coefficient and the structural resonance anomaly index are combined into a two-dimensional feature vector as the response matching feature; The vibration impact waveforms of the actual vibration response sequence at the start and stop transients of each wind-resistant adjustment action are extracted, and cross-correlation analysis is performed with the corresponding transient waveforms simulated by the digital twin to calculate the shape similarity and time lag of the waveforms. Based on the shape similarity and the time lag, an excitation conflict feature vector of the physical transmission mechanism is constructed.

[0010] In some embodiments, determining anomalies in the physical transmission mechanism based on the response matching features specifically includes: During the initial stage of normal commissioning and operation, response matching features under multiple wind resistance adjustment actions of different intensities were collected to obtain the baseline feature cloud map under healthy conditions. Calculate the Mahalanobis distance between the response matching feature vector under the current wind resistance adjustment action and the center of the baseline feature cloud map; If the Mahalanobis distance exceeds a preset abnormal threshold, the physical transmission mechanism is determined to be abnormal.

[0011] In some embodiments, locating the fault position of the physical transmission mechanism using the excitation conflict characteristics specifically includes: A fault mode feature library is established, which pre-stores standard excitation conflict feature vectors generated by simulation in a digital twin for different fault locations; Calculate the feature similarity between the excitation conflict feature vector and each standard feature vector in the fault mode feature library; Use all feature similarities to locate the fault location of the physical transmission mechanism.

[0012] Secondly, this application provides an adaptive wind-resistant pitched roof photovoltaic tile system, including a fault diagnosis unit, the fault diagnosis unit comprising: The data acquisition module is used to acquire the control command sequence of the drive device, the actual angle response sequence of the driven device, and the actual vibration response sequence of the driven device when the physical transmission mechanism performs adaptive wind resistance adjustment. The processing module is used to initialize the digital twin of the physical transmission mechanism, perform multibody dynamics prediction on the control command sequence based on the digital twin, and obtain the simulated angle response sequence and simulated vibration response sequence of the physical transmission mechanism in the current wind resistance adjustment action. The processing module is also used to perform time-frequency joint comparison of the simulated angle response sequence and the actual angle response sequence to obtain the transmission ratio offset coefficient of the physical transmission mechanism, and at the same time to evaluate the frequency band difference between the simulated vibration response sequence and the actual vibration response sequence to obtain the structural resonance anomaly index of the physical transmission mechanism. Then, the response matching characteristics and excitation conflict characteristics of the physical transmission mechanism are determined based on the transmission ratio offset coefficient and the structural resonance anomaly index. The execution module is used to determine the anomaly of the physical transmission mechanism based on the response matching features. When an anomaly is determined, the fault location of the physical transmission mechanism is located by the excitation conflict features.

[0013] Thirdly, this application provides a computer device, the computer device including a memory and a processor, the memory for storing a computer program, and the processor for calling and running the computer program from the memory, so that the computer device performs the above-described transmission mechanism fault diagnosis method.

[0014] Fourthly, this application provides a computer-readable storage medium storing instructions or code that, when executed on a computer, cause the computer to implement the aforementioned transmission mechanism fault diagnosis method.

[0015] The technical solutions provided by the embodiments disclosed in this application have the following beneficial effects: In the adaptive wind-resistant roof photovoltaic tile system and its transmission mechanism fault diagnosis method provided in this application, when the physical transmission mechanism performs adaptive wind-resistant adjustment action, the control command sequence of the driving device, the actual angle response sequence of the driven device, and the actual vibration response sequence are collected; a digital twin of the physical transmission mechanism is initialized, and multibody dynamics prediction is performed on the control command sequence based on the digital twin to obtain the simulated angle response sequence and simulated vibration response sequence of the physical transmission mechanism in the current wind-resistant adjustment action; the simulated angle response sequence and the actual angle response sequence are jointly compared in time and frequency to obtain the transmission ratio offset coefficient of the physical transmission mechanism, and the frequency band difference is evaluated on the simulated vibration response sequence and the actual vibration response sequence to obtain the structural resonance anomaly index of the physical transmission mechanism; then, the response matching characteristics and excitation conflict characteristics of the physical transmission mechanism are determined according to the transmission ratio offset coefficient and the structural resonance anomaly index; the physical transmission mechanism is judged for anomalies based on the response matching characteristics, and when an anomaly is determined, the fault location of the physical transmission mechanism is located through the excitation conflict characteristics.

[0016] Therefore, in this application, the physical transmission mechanism is judged for anomalies based on the response matching features. When an anomaly is determined, the fault location of the physical transmission mechanism is located through the excitation conflict features. First, by integrating the response matching features composed of the transmission ratio offset coefficient and the structural resonance anomaly index, a two-dimensional feature vector can be obtained that comprehensively quantifies the degree of deviation of the motion transmission accuracy and structural dynamic characteristics of the transmission mechanism from the health benchmark. The deep anomaly information of the angle response and vibration response in the time-frequency domain and frequency band energy domain is dimensionality reduced and fused, which can effectively characterize the overall health status of the transmission mechanism under continuous operation. This feature vector, serving as a high-order abstract measure of the consistency between the ideal response simulated by the digital twin and the actual system response at the steady-state level, provides a stable and robust criterion for subsequent anomaly judgment, avoiding the risk of misjudgment based on a single indicator and significantly enhancing the reliability of fault detection. Then, by extracting and analyzing the excitation conflict features constructed from the waveform and time differences of the actual and simulated vibration responses during the transient phase of wind-resistant adjustment, a set of feature vectors focusing on the degree of transient behavior mismatch under dynamic excitation can be obtained. This captures the transient impact response characteristics caused by faults such as gaps and stiffness abrupt changes. These characteristics are highly sensitive to the fault location. This feature vector, as a fingerprint information with strong directional characteristics for the fault location, can effectively distinguish similar abnormal manifestations caused by faults of different locations and types by matching it with a pre-built fault mode feature library, thus achieving a diagnostic leap from the existence of an anomaly to the location of the fault source. In summary, based on the above scheme, fault diagnosis based on digital twin virtual-real comparison can be realized, thereby improving the accuracy of fault detection and location in transmission mechanisms. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is an exemplary flowchart of a transmission mechanism fault diagnosis method according to some embodiments of this application; Figure 2 This is a flowchart illustrating the process of determining the structural resonance anomaly index according to some embodiments of this application; Figure 3 This is a schematic diagram of the structure of a fault diagnosis unit according to some embodiments of this application; Figure 4 This is a schematic diagram of the structure of a computer device that implements a fault diagnosis method for a transmission mechanism according to some embodiments of this application. Detailed Implementation

[0019] To better understand the technical solution of this application, the technical solution of this application will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0020] refer to Figure 1 The figure is an exemplary flowchart of a transmission mechanism fault diagnosis method according to some embodiments of this application. The transmission mechanism fault diagnosis method mainly includes the following steps: In step 101, when the physical transmission mechanism performs adaptive wind resistance adjustment, the control command sequence of the drive device, the actual angle response sequence of the driven device, and the actual vibration response sequence are collected.

[0021] In practice, the controller generates wind-resistant adjustment logic based on the feedback signal from the wind sensor and sends control commands containing the target position, speed, and acceleration to the drive device (usually a servo motor or stepper motor). These commands, arranged in chronological order, are recorded in their entirety to form a control command sequence for subsequent fault diagnosis and analysis. While the drive device executes this command sequence, a high-precision absolute encoder coaxially connected to the driven device's shaft continuously measures and records the real-time rotation angle of the driven device at a sampling frequency of no less than 10 times the control cycle, forming an actual angle response sequence that reflects its actual motion trajectory. At the same time, an industrial-grade triaxial accelerometer arranged in the transmission mechanism housing near the bearing seat or gear meshing area collects the vibration signal generated by the transmission mechanism during operation with a synchronous timestamp. After anti-aliasing filtering and analog-to-digital conversion, this signal yields an actual vibration response sequence that is time-synchronized with the actual angle response sequence.

[0022] It should be noted that, in this application, the control command sequence is a digital signal sequence used to describe the expected motion state of the drive device within a specific time interval; the actual angle response sequence is a measurement data sequence used to characterize the actual rotation angle change of the driven device under the drive of the control command; the actual vibration response sequence is a triaxial acceleration measurement data sequence used to reflect the dynamic characteristics and state of the mechanical structure of the transmission mechanism during motion; the high-precision encoder is a precision measuring element used to convert the angular displacement of the rotating shaft into a corresponding digital signal; the triaxial accelerometer is an electromechanical conversion device used to simultaneously measure the vibration acceleration in three orthogonal directions in space; and anti-aliasing filtering is a signal processing procedure used to limit the signal bandwidth before digitizing the analog signal to avoid spectral aliasing caused by high-frequency components.

[0023] In step 102, the digital twin of the physical transmission mechanism is initialized, and multibody dynamics prediction is performed on the control command sequence based on the digital twin to obtain the simulated angle response sequence and simulated vibration response sequence of the physical transmission mechanism in the current wind resistance adjustment action.

[0024] In some embodiments, the initialization of a digital twin of a physical transmission mechanism can be achieved as follows: First, obtain the complete design parameters of the physical transmission mechanism, including the number of teeth, module, pressure angle, center distance of gear pairs, diameter and length of the transmission shaft, bearing type and arrangement, and material properties and mass distribution information of each component. Second, based on multibody dynamics modeling, establish a virtual three-dimensional geometric model of the transmission mechanism in simulation software according to the above parameters, and assign kinematic pair constraints to the model (e.g., define gear meshing as a contact pair, and define bearing support as a rotary pair and elastic constraint). Then, calculate or set the contact stiffness and damping coefficient between each moving part, as well as the stiffness matrix of flexible bodies such as the transmission shaft, according to the material properties of the components, thereby constructing a parameterized dynamic equation that reflects the dynamic characteristics of the mechanism. Finally, associate the dynamic equation with the virtual geometric model through a solver to form a virtual entity that can receive drive commands and calculate the output dynamic response of the mechanism, i.e., a digital twin. The digital twin is a computable parameterized virtual model used to map and simulate the structure, motion, and dynamic behavior of the physical transmission mechanism in virtual space.

[0025] In some embodiments, the following steps can be used to perform multibody dynamics prediction on the control command sequence based on the digital twin to obtain the simulated angle response sequence and simulated vibration response sequence of the physical transmission mechanism in the current wind-resistant adjustment action: The control command sequence is converted into a time series of the equivalent input torque and speed of the drive unit; In the digital twin, based on the geometric parameters, inertial parameters, stiffness parameters and damping parameters of the transmission chain, a multi-degree-of-freedom dynamic equation is established, which includes gear pair meshing, bearing support and the flexibility of connecting parts. Using the equivalent input torque and rotational speed in the time series as boundary conditions, the dynamic equation is solved numerically with time steps, and the simulated angle response sequence and simulated vibration response sequence are output synchronously.

[0026] It should be noted that, in this application, the time series of equivalent input torque and rotational speed is a digital signal sequence used to drive the virtual model in the digital twin and corresponding to the actual output torque and speed change process of the physical drive device; the multi-degree-of-freedom dynamic equation is a set of differential equations describing the relationship between force, torque and acceleration of each component in the transmission mechanism in multiple independent motion directions; time-step numerical solution is a mathematical calculation method used to obtain approximate solutions of the dynamic equation at discrete time points through stepwise progressive calculation; the simulated angle response sequence is a sequence of calculated results characterizing the change of angular displacement of the driven device in the digital twin with time during simulated operation; and the simulated vibration response sequence is a sequence of calculated results characterizing the change of vibration acceleration of the preset monitoring point in the digital twin with time during simulated operation.

[0027] In specific implementation, firstly, converting the control command sequence into a time series of equivalent input torque and speed of the drive unit can be achieved in the following way: the control command sequence is processed, which includes the pulse width modulation signal or position command of the motor controller. By parsing the command and combining the known torque constant, moment of inertia and other characteristic parameters of the drive unit (e.g., servo motor), the torque magnitude and speed value on the output shaft of the drive unit at each moment are calculated. The set of torque values ​​and speed values ​​arranged in chronological order is taken as the time series of equivalent input torque and speed. Then, in the digital twin, based on the geometric parameters, inertia parameters, stiffness parameters and damping parameters of the transmission chain, a multi-degree-of-freedom dynamic equation including gear meshing, bearing support and connecting flexibility can be established in the following way: in the initialized digital twin, based on the geometric parameters (e.g., number of gear teeth), inertia parameters (e.g., moment of inertia of each rotating component), stiffness parameters (e.g., torsional stiffness of the shaft), and damping parameters (e.g., damping coefficient of the bearing) of the transmission chain, a dynamic model of the system is established. This model describes the system's motion using multiple generalized coordinates (e.g., the rotation angles of each gear) and employs mechanical principles such as the Lagrange equation or the Newton-Euler method to establish a set of multi-degree-of-freedom dynamic differential equations, including gear meshing forces, bearing support reactions, and elastic forces of connecting components. Finally, using the equivalent input torque and rotational speed in the time series as boundary conditions, the dynamic equations are solved numerically in time steps, and the simulated angle response sequence and simulated vibration response sequence are output synchronously. This can be achieved by using the Runge-Kutta method or other numerical integration methods, taking the time series of equivalent input torque and rotational speed as the system input, and solving the multi-degree-of-freedom dynamic equations step by step. At each solution time step, the simulated angular displacement (i.e., simulated angle response) of the driven unit (e.g., the output shaft) and the simulated vibration acceleration of the pre-set key monitoring points (e.g., the gearbox housing) in three directions are calculated. The continuous time step solution results are used as the simulated angle response sequence and simulated vibration response sequence.

[0028] In step 103, the simulated angle response sequence and the actual angle response sequence are compared in a time-frequency manner to obtain the transmission ratio offset coefficient of the physical transmission mechanism. At the same time, the simulated vibration response sequence and the actual vibration response sequence are evaluated for frequency band difference to obtain the structural resonance anomaly index of the physical transmission mechanism. Then, the response matching characteristics and excitation conflict characteristics of the physical transmission mechanism are determined based on the transmission ratio offset coefficient and the structural resonance anomaly index.

[0029] In some embodiments, the transmission ratio offset coefficient of the physical transmission mechanism can be obtained by performing a time-frequency joint comparison between the simulated angle response sequence and the actual angle response sequence. The simulated angle response sequence and the actual angle response sequence are synchronously resampled at equal intervals, and the simulated value and the actual value of the instantaneous transmission ratio are calculated respectively. In the time domain, the absolute difference between the actual and simulated values ​​of the instantaneous transmission ratio at each sampling moment is calculated to obtain the transmission ratio deviation sequence; In the frequency domain, a fast Fourier transform is performed on the transmission ratio deviation sequence to extract the peak value of the amplitude spectrum at the transmission characteristic frequency. The transmission ratio offset coefficient of the physical transmission mechanism is determined by weighting the root mean square value of the transmission ratio deviation sequence and the peak value of the amplitude spectrum at the characteristic frequency.

[0030] It should be noted that, in this application, the instantaneous transmission ratio simulated / actual value is a numerical value used to characterize the speed transmission relationship between the driving unit and the driven unit predicted by the digital twin / measured by the physical mechanism at a specific moment; the transmission ratio deviation sequence is a data sequence used to record the absolute difference between the instantaneous transmission ratio actual value and the simulated value in chronological order; the Fast Fourier Transform is a classical mathematical algorithm used to convert a time-domain signal into a frequency-domain signal to analyze its frequency components and their intensity; the transmission characteristic frequency is a theoretically calculated frequency value corresponding to the periodic occurrence rate of a specific physical event (e.g., single tooth meshing) in the transmission mechanism; the amplitude spectrum peak value is a numerical value used to represent the signal energy strength at a specific frequency on the signal spectrum diagram; and the transmission ratio offset coefficient is a single numerical index used to comprehensively quantify the degree of deviation of the actual motion transmission relationship of the transmission mechanism from the prediction of the ideal model of the digital twin.

[0031] In specific implementation, firstly, the simulated angle response sequence and the actual angle response sequence are synchronously resampled at equal intervals, and the simulated and actual instantaneous transmission ratio values ​​are calculated respectively. This can be achieved in the following way: In order to ensure that the simulated angle response sequence and the actual angle response sequence are accurately aligned on the time axis for subsequent comparison, the two sequences are resampled using the same sampling time interval. This process is called synchronous equal-interval resampling. After resampling, for each sampling moment, the simulated angle response value, the actual angle response value, and the known nominal theoretical transmission ratio from the driving unit to the driven unit are used to calculate the simulated and actual instantaneous transmission ratio values ​​at that moment through a simple proportional calculation relationship. The instantaneous transmission ratio can be approximated by the differential of the rotation angle of the driving unit with respect to the rotation angle of the driven unit (i.e., the angular velocity ratio), or by the ratio of angle increments within a fixed time window. Secondly, in the time domain, the absolute difference between the actual and simulated instantaneous transmission ratio values ​​at each sampling moment is calculated to obtain the transmission ratio deviation sequence. This can be achieved by subtracting the simulated instantaneous transmission ratio value from the actual instantaneous transmission ratio value at each sampling moment and taking the absolute value, resulting in a transmission ratio deviation sequence that varies with time, directly reflecting the fluctuation of transmission accuracy over time. Then, in the frequency domain, a fast Fourier transform is performed on the transmission ratio deviation sequence to extract the peak amplitude spectrum at the transmission characteristic frequency. This can be achieved by applying a fast Fourier transform to the transmission ratio deviation sequence, converting it from the time domain to the frequency domain. A spectrum diagram is obtained. From this spectrum diagram, specific frequencies directly related to the physical characteristics of the transmission mechanism (e.g., gear meshing frequency and its harmonics) are identified and extracted, namely the transmission characteristic frequencies. The peak values ​​of the amplitude spectrum at these characteristic frequency points are recorded. Finally, the transmission ratio offset coefficient of the physical transmission mechanism can be determined by the following method based on the root mean square value of the transmission ratio deviation sequence and the weighted sum of the peak values ​​of the amplitude spectrum at the characteristic frequencies: the root mean square value of the transmission ratio deviation sequence in the time domain is calculated to measure the amplitude level of the overall deviation; at the same time, the peak values ​​of the amplitude spectrum at each transmission characteristic frequency extracted in the frequency domain are weighted and summed, and the characteristic frequencies that are more related to the fault mechanism are given higher weights. The root mean square value and the value of the weighted sum are fused and calculated through a normalization formula to obtain a scalar value between 0 and 1 as the transmission ratio offset coefficient.

[0032] In some embodiments, the frequency band difference between the simulated vibration response sequence and the actual vibration response sequence is evaluated to obtain the structural resonance anomaly index of the physical transmission mechanism, with reference to... Figure 2 The figure described above is a flowchart illustrating the determination of the structural resonance anomaly index in some embodiments of this application. In this embodiment, the determination of the structural resonance anomaly index can be achieved using the following steps: In step 1031, wavelet packet decomposition is performed on the simulated vibration response sequence and the actual vibration response sequence respectively, and the decomposition is performed to the preset frequency band level. In step 1032, the energy ratio of the actual vibration signal to the simulated vibration signal in each corresponding frequency band of the frequency band hierarchy is calculated, thereby obtaining the energy logarithmic ratio of each frequency band in the frequency band hierarchy; In step 1033, the structural resonance anomaly index of the physical transmission mechanism is determined by all energy logarithmic ratios.

[0033] It should be noted that in this application, wavelet packet decomposition is used to recursively subdivide the signal in both high-frequency and low-frequency components, thereby providing a signal processing method with more refined and flexible frequency band division than traditional wavelet decomposition; the preset frequency band level is a parameter used to specify the recursive depth of wavelet packet decomposition, thereby determining the number and bandwidth of the final signal divided into sub-bands; the energy logarithmic ratio is a value used to quantify the degree of difference between the energy of the actual vibration signal and the energy of the simulated vibration signal within a specific frequency band, and is logarithmically compressed; the structural resonance anomaly index is a normalized scalar index used to comprehensively evaluate the degree of deviation of the current structural dynamic characteristics of the physical transmission mechanism from the state of the digital twin health model.

[0034] In specific implementation, firstly, wavelet packet decomposition is performed on both the simulated and actual vibration response sequences. Decomposition to a preset frequency band level can be achieved as follows: using the wavelet packet decomposition algorithm, multi-level decomposition is performed on both the simulated and actual vibration response sequences, selecting a suitable wavelet basis function (e.g., Daubechies wavelet) and a preset number of decomposition levels. This number of levels determines the fineness of the frequency band division. The decomposition process recursively decomposes each original vibration signal into a series of sub-band signals covering different frequency ranges. These sub-bands together form a complete tree structure until the preset frequency band level is reached. At this point, the frequency band of the original signal is uniformly and finely divided into several preset number of sub-bands. Then, the energy ratio of the actual vibration signal to the simulated vibration signal within each corresponding frequency band in the frequency band level is calculated. The logarithmic energy ratio of each frequency band in the frequency band level can be obtained as follows: for each corresponding sub-band obtained after wavelet packet decomposition, the energy of the actual vibration signal and the energy of the simulated vibration signal within that band are calculated. Specifically, first, the sub-band signal coefficients corresponding to the frequency band are extracted, and then the sum of the squares of these coefficients is calculated as the energy of the frequency band. Next, the energy value of the actual vibration signal in that frequency band is divided by the energy value of the simulated vibration signal in the same frequency band to obtain an energy ratio. To avoid excessively large numerical ranges and to enhance sensitivity to small differences, the energy ratio is taken as its logarithm to the natural constant e, yielding the energy logarithmic ratio of the frequency band. This calculation is repeated for all decomposed sub-bands to obtain a set containing the energy logarithmic ratios of all frequency bands. Finally, the structural resonance anomaly index of the physical transmission mechanism can be determined using all energy logarithmic ratios in the following way: the structural resonance anomaly index is determined based on the energy logarithmic ratios of all frequency bands. Typically, when the health status of a transmission mechanism changes, its structural dynamic characteristics (e.g., natural frequency, damping) alter, leading to significant differences in the energy distribution of certain frequency bands (usually those related to the structure's natural frequency or transmission characteristic frequency). Therefore, from all energy logarithmic ratios, those frequency bands highly correlated with the known structural resonant frequency (called critical resonant frequency bands) can be identified, and the absolute values ​​of the energy logarithmic ratios of these critical frequency bands can be calculated. These absolute values ​​are then weighted and averaged, with weights pre-set based on the sensitivity of each frequency band to faults. To obtain a standardized and easily interpretable index, the weighted average can be normalized, for example, mapped to a range of 0 to 1, where 0 represents perfect consistency with the simulated state and 1 represents the maximum abnormal deviation. This final normalized value is the structural resonance anomaly index.

[0035] In some embodiments, determining the response matching characteristics and excitation conflict characteristics of the physical transmission mechanism based on the transmission ratio offset coefficient and the structural resonance anomaly index can be achieved through the following steps: The transmission ratio offset coefficient and the structural resonance anomaly index are combined into a two-dimensional feature vector as the response matching feature; The vibration impact waveforms of the actual vibration response sequence at the start and stop transients of each wind-resistant adjustment action are extracted, and cross-correlation analysis is performed with the corresponding transient waveforms simulated by the digital twin to calculate the shape similarity and time lag of the waveforms. Based on the shape similarity and the time lag, an excitation conflict feature vector of the physical transmission mechanism is constructed.

[0036] It should be noted that, in this application, the response matching feature is a feature vector characterizing the degree of agreement between the motion transmission accuracy and structural vibration characteristics of the physical transmission mechanism under continuous operation and the ideal prediction of the digital twin; the start-up and stop transients refer to the time intervals of the brief non-stationary processes caused by sudden changes in state at the start and end of the wind-resistant adjustment action of the transmission mechanism; the vibration shock waveform describes the specific vibration signal morphology caused by mechanical shock excitation during the start-up or stop transients, characterized by drastic amplitude changes and short duration; cross-correlation analysis is a mathematical method used to calculate the similarity of two signals at different time offsets and extract the maximum similarity value and its corresponding time difference (hysteresis); shape similarity is a numerical index used to quantify the consistency of the amplitude change morphology of two waveform signals; time hysteresis is a numerical value used to represent the overall offset of one waveform signal relative to another waveform signal on the time axis; and the excitation conflict feature vector is a feature vector characterizing the difference in waveform and time synchronization between the transient dynamic response of the physical transmission mechanism and the ideal transient response of the digital twin when subjected to sudden excitation by external control commands.

[0037] In specific implementation, firstly, combining the transmission ratio offset coefficient and the structural resonance anomaly index into a two-dimensional feature vector as a response matching feature can be achieved in the following way: The two scalar indices, the transmission ratio offset coefficient and the structural resonance anomaly index, are combined in a preset order (e.g., the transmission ratio offset coefficient as the first dimension and the structural resonance anomaly index as the second dimension) to form an array with two elements. This two-dimensional array, i.e., the two-dimensional feature vector, is defined as the response matching feature of the physical transmission mechanism. This feature vector comprehensively characterizes the degree of matching between the overall response of the physical mechanism and the ideal prediction of the digital twin during continuous operation from two dimensions: macroscopic motion transmission accuracy and microscopic structural dynamic characteristics. Then, the vibration impact waveforms of the actual vibration response sequence at the start and stop transients of each wind-resistant adjustment action are extracted, and cross-correlation analysis is performed with the corresponding transient waveforms simulated by the digital twin. The waveform shape similarity and time lag can be calculated in the following way: For each independent wind-resistant adjustment action (the action cycle starts from the issuance of the start command and ends when the stop command is completed and the mechanism reaches a steady state), data from two key transient stages are extracted from the records synchronized with the actual vibration response sequence: one is the short-time vibration signal segment at the moment of start of the adjustment action, and the other is the short-time vibration signal segment at the moment of stop of the adjustment action. The two signal segments contain impact vibration waveforms caused by sudden load changes, gear backlash reversal, etc. Similarly, the corresponding simulated vibration waveforms at the start and stop transients are extracted from the simulated vibration response sequence according to the same control command time marker. Then, cross-correlation analysis is performed on each pair of corresponding actual transient waveforms and simulated transient waveforms (e.g., the actual waveform and simulated waveform of the startup transient). The cross-correlation analysis calculates the sum of the dot products of the two waveforms at different relative time offsets to obtain a cross-correlation function sequence. The maximum value can be found from this cross-correlation function sequence, and the normalization coefficient corresponding to the maximum value reflects the shape similarity of the two waveforms under optimal alignment. At the same time, the difference between the position index of the maximum value in the cross-correlation function sequence and the zero offset, after sampling time conversion, yields the time lag between the two waveforms, which reflects the time delay or advance of the physical mechanism response relative to the ideal model prediction. Finally, based on the shape similarity and the time lag, the excitation conflict feature vector of the physical transmission mechanism can be constructed in the following way: using the shape similarity and time lag of all transient waveforms to construct the excitation conflict feature vector.For example, four indicators can be extracted: shape similarity of the starting transient, time lag of the starting transient, shape similarity of the stopping transient, and time lag of the stopping transient. These indicators are arranged in a predetermined order to form a four-dimensional feature vector. This vector is the excitation conflict feature vector of the physical transmission mechanism. This vector focuses on the difference between the actual dynamic response of the system and the ideal model in waveform shape and time synchronization when the system is subjected to a sudden change (excitation) in the control command. This reveals the impact of potential faults (such as excessive clearance, decreased stiffness, and abnormal damping) on ​​the transient characteristics of the system.

[0038] In step 104, the physical transmission mechanism is judged to be abnormal based on the response matching feature. When an abnormality is judged, the fault location of the physical transmission mechanism is located by the excitation conflict feature.

[0039] In some embodiments, the anomaly determination of the physical transmission mechanism based on the response matching characteristics can be achieved by the following steps: During the initial stage of normal commissioning and operation, response matching features under multiple wind resistance adjustment actions of different intensities were collected to obtain the baseline feature cloud map under healthy conditions. Calculate the Mahalanobis distance between the response matching feature vector under the current wind resistance adjustment action and the center of the baseline feature cloud map; If the Mahalanobis distance exceeds a preset abnormal threshold, the physical transmission mechanism is determined to be abnormal.

[0040] It should be noted that, in this application, the baseline feature cloud map is a set of two-dimensional data points and its statistical representation used to describe the normal fluctuation range and distribution pattern of the response matching feature vector of the transmission mechanism under fault-free conditions in the feature space; the Mahalanobis distance is a statistical distance metric used to measure the distance between a data point (current feature vector) and the distribution center of a data set (baseline feature cloud map), which takes into account the correlation and variance between the dimensions of the dataset; the preset anomaly threshold is used to compare with the Mahalanobis distance to determine whether the current state deviates significantly from the preset or critical value of the health baseline.

[0041] In specific implementation, firstly, during the initial stage of normal commissioning and operation, response matching characteristics under multiple wind resistance adjustment actions of different intensities are collected to obtain the baseline feature cloud map under healthy conditions. This can be achieved in the following way: during the initial stage after the transmission mechanism has been installed and commissioned and confirmed to be fault-free, or during a stable observation period after the system has been put into normal operation, a series of wind resistance adjustment actions of different intensities are executed in a planned manner. These actions should cover the typical operating range expected by the system, for example, from minimum angle adjustment to maximum angle adjustment, and executed at different speed levels. When each adjustment action is executed, the corresponding response matching feature vector (i.e., a two-dimensional vector composed of the transmission ratio offset coefficient and the structural resonance anomaly index) is collected and calculated simultaneously according to the method defined in claims 1 to 5. All these response matching feature vectors collected under healthy conditions are gathered together to form a set of two-dimensional data points. This set exhibits a certain distribution pattern in the two-dimensional feature space spanned by the transmission ratio offset coefficient and the structural resonance anomaly index. This distribution pattern and its core area are defined as the baseline feature cloud map under healthy conditions. Typically, the center (or centroid) of the cloud map can be obtained by calculating the arithmetic mean of all data points in each feature dimension. Then, the Mahalanobis distance between the response matching feature vector under the current wind resistance adjustment action and the center of the benchmark feature cloud map can be calculated as follows: when it is necessary to diagnose the current operating state, for the current wind resistance adjustment action, a current response matching feature vector is calculated according to the same process. Then, the Mahalanobis distance between the current vector and the aforementioned benchmark feature cloud map center is calculated. The calculation of the Mahalanobis distance not only considers the coordinate difference between the current point and the center point of the cloud map in each dimension, but also considers the shape and correlation of the distribution of each data point in the benchmark feature cloud map. Specifically, it is achieved by introducing the inverse matrix of the covariance matrix of the benchmark feature cloud map. This distance value is a scalar that quantifies the degree of deviation of the current state point from the overall distribution of healthy states, while also considering the inherent variability of health data. Finally, if the Mahalanobis distance exceeds a preset anomaly threshold, the physical transmission mechanism is deemed to be abnormal. This can be achieved by comparing the Mahalanobis distance value with a preset anomaly threshold. This anomaly threshold is preset based on the statistical characteristics of the health baseline feature cloud map data; for example, it can be set as the 99th percentile of the Mahalanobis distance distribution of healthy states, or a multiple determined empirically. If the calculated Mahalanobis distance exceeds the preset anomaly threshold, the overall response characteristics exhibited by the physical transmission mechanism in the current adjustment action are determined to have significantly deviated from the health baseline, i.e., an anomaly exists.

[0042] In some embodiments, locating the fault position of the physical transmission mechanism using the excitation conflict characteristics can be achieved through the following steps: A fault mode feature library is established, which pre-stores standard excitation conflict feature vectors generated by simulation in a digital twin for different fault locations; Calculate the feature similarity between the excitation conflict feature vector and each standard feature vector in the fault mode feature library; Use all feature similarities to locate the fault location of the physical transmission mechanism.

[0043] It should be noted that, in this application, the fault mode feature library is a database used to store standard excitation conflict feature vectors with clear fault labels and their related metadata generated by digital twins under preset fault parameters of different types and locations; the standard excitation conflict feature vector is a feature vector used to represent the typical transient response characteristics exhibited by a certain fault mode (including location and type information) under ideal simulation conditions; feature similarity is a numerical index used to quantify the degree of closeness or morphological similarity between two excitation conflict feature vectors in mathematical space; fault location is the identification information used to refer to the specific physical component or subsystem in the transmission mechanism that has an abnormality or damage.

[0044] In practical implementation, firstly, a fault mode feature library is established, which pre-stores standard excitation conflict feature vectors generated in the digital twin for different fault locations. This can be achieved by pre-constructing a fault mode feature library. This feature library is established by artificially and selectively introducing various known and typical local fault parameter variations into the digital twin model. Specifically, for different fault locations requiring diagnosis, such as wear of the drive-side gear, damage of the driven-side bearing, misalignment of the intermediate drive shaft, and excessive coupling clearance, the parameters of the corresponding component model in the digital twin are modified. For gear wear, the meshing stiffness of the gear teeth can be reduced; for bearing damage, asymmetry can be introduced into the support stiffness matrix of the corresponding bearing or its value can be reduced; for misalignment, the installation angle offset parameter of the shaft system can be adjusted. Then, the modified digital twin model is driven to perform a series of standard wind resistance adjustment actions identical to those acquired during health status acquisition, and the excitation conflict feature vector corresponding to the specific fault mode is calculated according to the method for generating excitation conflict feature vectors. Each combination of fault location and severity will generate one or more standard excitation conflict feature vectors, which are stored in the feature library along with their corresponding fault type labels (e.g., "moderate wear of drive-side gears", "damage to the outer ring of load-side bearings"). Then, the feature similarity between the excitation conflict feature vector and each standard feature vector in the fault mode feature library can be calculated in the following way: when the actual physical transmission mechanism is determined to be abnormal, the actual excitation conflict feature vector extracted according to the aforementioned steps needs to be compared with each standard feature vector in the fault mode feature library. The comparison is achieved by calculating feature similarity. Euclidean distance or cosine similarity measurement methods are usually used. Taking Euclidean distance as an example, the sum of squared differences between the actual vector and each standard vector in the feature library in each dimension is calculated, and the square root is taken to obtain a distance value. For ease of comparison and interpretation, this distance value can be converted into a similarity score; for example, the smaller the distance, the higher the similarity. This yields a set of feature similarity values ​​between the actual feature vector and each standard fault mode feature vector in the feature library. Finally, locating the fault position of the physical transmission mechanism using all feature similarities can be achieved by: selecting the standard feature vector with the highest feature similarity to the actual feature vector to be diagnosed, and outputting the fault position label corresponding to this standard feature vector as the diagnostic result—that is, the nearest neighbor classification method. A more robust strategy can use the K-nearest neighbor algorithm, which finds the K standard feature vectors with the highest feature similarity, examines the fault position labels corresponding to these K neighbor vectors, and uses voting or weighted voting (with the weight being their respective similarity) to obtain the final fault position diagnosis result.

[0045] Furthermore, in another aspect of this application, in some embodiments, this application provides an adaptive wind-resistant pitched roof photovoltaic tile system, which includes a fault diagnosis unit, referencing... Figure 3 The figure is a schematic diagram of the structure of a fault diagnosis unit according to some embodiments of this application. The fault diagnosis unit includes: a data acquisition module 201, a processing module 202, and an execution module 203, which are described below: The acquisition module 201 in this application is mainly used to acquire the control command sequence of the driving device, the actual angle response sequence of the driven device, and the actual vibration response sequence of the driven device when the physical transmission mechanism performs adaptive wind resistance adjustment action. Processing module 202, in this application, is used to initialize the digital twin of the physical transmission mechanism, perform multibody dynamics prediction on the control command sequence based on the digital twin, and obtain the simulated angle response sequence and simulated vibration response sequence of the physical transmission mechanism in the current wind resistance adjustment action; It should be noted that the processing module 202 is also used to perform time-frequency joint comparison of the simulated angle response sequence and the actual angle response sequence to obtain the transmission ratio offset coefficient of the physical transmission mechanism, and at the same time to evaluate the frequency band difference between the simulated vibration response sequence and the actual vibration response sequence to obtain the structural resonance anomaly index of the physical transmission mechanism, and then determine the response matching characteristics and excitation conflict characteristics of the physical transmission mechanism based on the transmission ratio offset coefficient and the structural resonance anomaly index. The execution module 203 in this application is mainly used to make anomaly judgment on the physical transmission mechanism based on the response matching features. When an anomaly is determined, the fault location of the physical transmission mechanism is located by the excitation conflict features.

[0046] The foregoing detailed examples of the adaptive wind-resistant pitched roof photovoltaic tile system and its transmission mechanism fault diagnosis method provided in the embodiments of this application. It is understood that the corresponding device, in order to achieve the above functions, includes hardware structures and / or software modules corresponding to the execution of each function. Those skilled in the art should readily recognize that, in conjunction with the units and algorithm steps of the various examples described in the embodiments disclosed herein, this application can be implemented in hardware or a combination of hardware and computer software. Whether a function is executed in hardware or by computer software driving hardware depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0047] In some embodiments, this application also provides a computer device, the computer device including a memory and a processor, the memory for storing a computer program, and the processor for calling and running the computer program from the memory, so that the computer device performs the above-described transmission mechanism fault diagnosis method.

[0048] In some embodiments, reference Figure 4 The dashed lines in the figure indicate that the unit or module is optional. This figure is a structural schematic diagram of a computer device for implementing a transmission mechanism fault diagnosis method according to an embodiment of this application. The transmission mechanism fault diagnosis method described in the above embodiments can... Figure 4 The computer device shown is used to implement this, and the computer device includes at least one processor 301, a memory 302 and at least one communication unit 305. The computer device may be a terminal device, a server or a chip.

[0049] Processor 301 can be a general-purpose processor or a special-purpose processor. For example, processor 301 can be a central processing unit (CPU), which can be used to control computer devices, execute software programs, and process data from software programs. The computer device may also include a communication unit 305 for inputting (receiving) and outputting (transmitting) signals.

[0050] For example, the computer device may be a chip, and the communication unit 305 may be the input and / or output circuit of the chip, or the communication unit 305 may be the communication interface of the chip, which may be a component of a terminal device, network device or other device.

[0051] For example, the computer device may be a terminal device or a server, and the communication unit 305 may be a transceiver of the terminal device or the server, or the communication unit 305 may be a transceiver circuit of the terminal device or the server.

[0052] The computer device may include one or more memories 302 storing a program 304. The program 304 can be executed by a processor 301 to generate instructions 303, causing the processor 301 to execute the method described in the above method embodiments according to the instructions 303. Optionally, the memory 302 may also store data (such as a target audit model). Optionally, the processor 301 may also read data stored in the memory 302, which may be stored at the same storage address as the program 304, or it may be stored at a different storage address than the program 304.

[0053] The processor 301 and memory 302 can be configured separately or integrated together, for example, integrated on the system on chip (SOC) of the terminal device.

[0054] It should be understood that each step of the above method embodiment can be completed by hardware logic circuits or software instructions in the processor 301. The processor 301 can be a CPU, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, such as discrete gates, transistor logic devices, or discrete hardware components.

[0055] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0056] For example, in some embodiments, this application also provides a computer-readable storage medium storing instructions or code that, when executed on a computer, cause the computer to implement the above-described transmission mechanism fault diagnosis method.

[0057] Although preferred embodiments of this application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of this application.

[0058] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of the claims of this application and their equivalents, this application also intends to include such modifications and variations.

Claims

1. A method for diagnosing transmission mechanism faults, used in an adaptive wind-resistant roof photovoltaic tile system for diagnosing transmission mechanism faults, characterized in that, Includes the following steps: When the physical transmission mechanism performs adaptive wind resistance adjustment, the control command sequence of the drive device, the actual angle response sequence of the driven device, and the actual vibration response sequence are collected. Initialize the digital twin of the physical transmission mechanism, and perform multibody dynamics prediction on the control command sequence based on the digital twin to obtain the simulated angle response sequence and simulated vibration response sequence of the physical transmission mechanism in the current wind resistance adjustment action; The simulated angle response sequence and the actual angle response sequence are compared in time and frequency to obtain the transmission ratio offset coefficient of the physical transmission mechanism. At the same time, the simulated vibration response sequence and the actual vibration response sequence are evaluated for frequency band difference to obtain the structural resonance anomaly index of the physical transmission mechanism. Then, the response matching characteristics and excitation conflict characteristics of the physical transmission mechanism are determined based on the transmission ratio offset coefficient and the structural resonance anomaly index. Based on the response matching features, anomalies are determined in the physical transmission mechanism. When an anomaly is determined, the fault location of the physical transmission mechanism is located through the excitation conflict features.

2. The method as described in claim 1, characterized in that, Based on the digital twin, multibody dynamics prediction is performed on the control command sequence to obtain the simulated angle response sequence and simulated vibration response sequence of the physical transmission mechanism in the current wind-resistant adjustment action. Specifically, these include: The control command sequence is converted into a time series of the equivalent input torque and speed of the drive unit; In the digital twin, based on the geometric parameters, inertial parameters, stiffness parameters and damping parameters of the transmission chain, a multi-degree-of-freedom dynamic equation is established, which includes gear pair meshing, bearing support and the flexibility of connecting parts. Using the equivalent input torque and rotational speed in the time series as boundary conditions, the dynamic equation is solved numerically with time steps, and the simulated angle response sequence and simulated vibration response sequence are output synchronously.

3. The method as described in claim 1, characterized in that, The transmission ratio offset coefficient of the physical transmission mechanism is obtained by performing a time-frequency joint comparison between the simulated angle response sequence and the actual angle response sequence. The simulated angle response sequence and the actual angle response sequence are synchronously resampled at equal intervals, and the simulated value and the actual value of the instantaneous transmission ratio are calculated respectively. In the time domain, the absolute difference between the actual and simulated values ​​of the instantaneous transmission ratio at each sampling moment is calculated to obtain the transmission ratio deviation sequence; In the frequency domain, a fast Fourier transform is performed on the transmission ratio deviation sequence to extract the peak value of the amplitude spectrum at the transmission characteristic frequency. The transmission ratio offset coefficient of the physical transmission mechanism is determined by weighting the root mean square value of the transmission ratio deviation sequence and the peak value of the amplitude spectrum at the characteristic frequency.

4. The method as described in claim 1, characterized in that, The frequency band difference between the simulated vibration response sequence and the actual vibration response sequence is evaluated to obtain the structural resonance anomaly index of the physical transmission mechanism, which specifically includes: Wavelet packet decomposition is performed on both the simulated vibration response sequence and the actual vibration response sequence, decomposing them to a preset frequency band level. Calculate the energy ratio of the actual vibration signal to the simulated vibration signal in each corresponding frequency band of the frequency band hierarchy, and then obtain the energy logarithmic ratio of each frequency band in the frequency band hierarchy; The structural resonance anomaly index of the physical transmission mechanism is determined by all energy logarithmic ratios.

5. The method as described in claim 1, characterized in that, The determination of the response matching characteristics and excitation conflict characteristics of the physical transmission mechanism based on the transmission ratio offset coefficient and the structural resonance anomaly index specifically includes: The transmission ratio offset coefficient and the structural resonance anomaly index are combined into a two-dimensional feature vector as the response matching feature; The vibration impact waveforms of the actual vibration response sequence at the start and stop transients of each wind-resistant adjustment action are extracted, and cross-correlation analysis is performed with the corresponding transient waveforms simulated by the digital twin to calculate the shape similarity and time lag of the waveforms. Based on the shape similarity and the time lag, an excitation conflict feature vector of the physical transmission mechanism is constructed.

6. The method as described in claim 1, characterized in that, The anomaly detection of the physical transmission mechanism based on the aforementioned response matching features specifically includes: During the initial stage of normal commissioning and operation, response matching features under multiple wind resistance adjustment actions of different intensities were collected to obtain the baseline feature cloud map under healthy conditions. Calculate the Mahalanobis distance between the response matching feature vector under the current wind resistance adjustment action and the center of the baseline feature cloud map; If the Mahalanobis distance exceeds a preset abnormal threshold, the physical transmission mechanism is determined to be abnormal.

7. The method as described in claim 1, characterized in that, Locating the fault location of the physical transmission mechanism using the aforementioned excitation conflict characteristics specifically includes: A fault mode feature library is established, which pre-stores standard excitation conflict feature vectors generated by simulation in a digital twin for different fault locations; Calculate the feature similarity between the excitation conflict feature vector and each standard feature vector in the fault mode feature library; Use all feature similarities to locate the fault location of the physical transmission mechanism.

8. An adaptive wind-resistant pitched roof photovoltaic tile system, the adaptive wind-resistant pitched roof photovoltaic tile system including a fault diagnosis unit, characterized in that, The fault diagnosis unit includes: The data acquisition module is used to acquire the control command sequence of the drive device, the actual angle response sequence of the driven device, and the actual vibration response sequence of the driven device when the physical transmission mechanism performs adaptive wind resistance adjustment. The processing module is used to initialize the digital twin of the physical transmission mechanism, perform multibody dynamics prediction on the control command sequence based on the digital twin, and obtain the simulated angle response sequence and simulated vibration response sequence of the physical transmission mechanism in the current wind resistance adjustment action. The processing module is also used to perform time-frequency joint comparison of the simulated angle response sequence and the actual angle response sequence to obtain the transmission ratio offset coefficient of the physical transmission mechanism, and at the same time to evaluate the frequency band difference between the simulated vibration response sequence and the actual vibration response sequence to obtain the structural resonance anomaly index of the physical transmission mechanism. Then, the response matching characteristics and excitation conflict characteristics of the physical transmission mechanism are determined based on the transmission ratio offset coefficient and the structural resonance anomaly index. The execution module is used to determine the anomaly of the physical transmission mechanism based on the response matching features. When an anomaly is determined, the fault location of the physical transmission mechanism is located by the excitation conflict features.

9. A computer device, characterized in that, The computer device includes a memory and a processor, the memory being used to store computer programs, and the processor being used to call and run the computer programs from the memory, causing the computer device to perform the transmission mechanism fault diagnosis method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores instructions or code that, when executed on a computer, cause the computer to implement the transmission mechanism fault diagnosis method as described in any one of claims 1 to 7.

Images