A device comfort measurement fault identification method, system and device

Abstract

The application belongs to the field of data processing, and provides a device comfort degree measurement fault identification method, system and device. A plurality of modal sequences consistent with the length of the sampling signal segment and aligned in bit sequence are obtained through variational mode decomposition. For each sampling signal segment, bit sequence is traversed to form a signal segment extended wave. The peak value of the signal segment extended wave is extracted, and the features of the peak value and its bit sequence are used to generate an extended wave ratio sequence. The difference between the wave ratio sorting of each sampling signal segment and its original sorting is used as the wave ratio drop of the sampling signal segment. The median of the wave ratio drop of all sampling signal segments is used to screen each sampling signal segment according to the wave ratio drop, and a signal comfort abnormal segment is identified. The amplitude fluctuation of random noise is not sensitive, but the envelope impact and bit sequence shift of a specific frequency band are more sensitive, which can prompt slight problems such as guide shoe gap, guide rail joint impact, traction machine transmission abnormality, etc. in the early stage, and is beneficial to preventive maintenance.

Description

Technical Field

[0001] This invention belongs to the field of data processing, and specifically relates to a method, system and device for identifying faults in equipment comfort measurement. Background Technology

[0002] Automated displacement equipment, as a core transportation tool in modern buildings, refers to various mechanical devices capable of performing at least one complete movement along a fixed path, such as autonomous vehicles, automated trams, automated elevators, escalators, automated lifts, automated unmanned aerial vehicles, and automated cable cars. Their operational safety, stability, and passenger comfort are becoming key concerns for both regulators and property owners. A typical detection chain involves sensing and collecting signals such as acceleration, angular velocity, position, and / or velocity. Signal preprocessing, such as attitude correction and noise reduction, is then performed. The system then segments and identifies signal segments for starting, accelerating, constant speed, deceleration, leveling adjustments, and stopping. Finally, performance evaluations are conducted, particularly assessing mobility comfort, vibration intensity, impact characteristics, and fault / anomaly detection, followed by maintenance recommendations and closed-loop verification. With the implementation of predictive maintenance concepts, the industry is gradually moving from offline spot checks to online monitoring and intelligent diagnostics. However, there is still significant room for improvement in engineering feasibility and consistency, particularly in areas such as stable segmentation under complex operating conditions, robust feature extraction of multimodal signals, and interpretable mapping of indicators to passenger experience.

[0003] Existing technologies suffer from insufficient robustness in segmentation, often relying on speed, time thresholds, or fixed thresholds. This makes it difficult to stably reproduce stage boundaries under varying loads, speed curves, and guide rail conditions. Short stages, such as leveling and fine-tuning, are easily segmented or mis-segmented. Previous practices primarily relied on black-box scoring using full-segment spectrum thresholds, peak-to-peak values, or fuzzy neural networks, which are insensitive to changes in positional structure and have poor cross-device comparability. Currently, many so-called intelligent control models have fixed parameters, leading to unstable decomposition. When switching between broadband impulses and steady-state narrowband, fixed decomposition parameters, such as fixed mode numbers or bandwidth constraints, easily cause mixed modes or over-decomposition, resulting in misjudgments. Furthermore, many solutions stop at alarm scoring, lacking a rule base that systematically maps technical characteristics such as frequency band, envelope impulses, and harmonic structure to component categories, maintenance actions, and re-inspection plans, making closed-loop implementation difficult. For example, patent document CN105836560A provides an elevator comfort testing system that can perform comfort calculation and optimization control based on fuzzy artificial neural networks, complete data acquisition, signal conditioning, and comfort calculation, and can also be linked to control and improve comfort. However, it lacks segmentation and adaptive operation condition, focuses on overall comfort modeling, and does not show adaptive identification and segmented evaluation criteria for the boundaries of stages such as start-up, constant speed, and leveling. Moreover, the interpretability of features is weak, and its explanatory power for impact location, positional structure, and specific frequency band anomalies is limited, which is not conducive to consistent reuse across equipment.

[0004] For example, patent document CN112320520B proposes a method for detecting abnormal elevator vibrations based on residual analysis. It identifies abnormal vibrations through residual decomposition, peak-to-peak value anomaly detection, and stage division, obtaining information such as vibration duration, intensity, and occurrence stage. It criticizes methods that rely solely on acceleration thresholds or require a large number of prior samples. However, this method uses the peak-to-peak value of the residual sequence as the primary anomaly, lacking sufficient analysis of frequency band assignment and envelope impact structure. This method also has limited characterization of intra-segment structures; although it distinguishes stages, it does not introduce positional structural indicators to measure the location and cumulative effect of impacts within a segment. Furthermore, patent document CN109264521B mentions an elevator fault diagnosis device that detects the car's position and / or speed and performs spectral analysis. It triggers an alarm when the amplitude of a certain frequency exceeds a set threshold, emphasizing its simplicity and low cost. However, it uses a fixed threshold and is single-channel dependent, with the spectral amplitude threshold as its core. This makes it difficult to adapt to different loads, velocity curves, and noise conditions, resulting in weak cross-scene transferability. Furthermore, it does not integrate multi-modal information such as acceleration and attitude correction, making it difficult to distinguish between inherent large amplitude acceleration / deceleration and abnormal impacts. Additionally, this method's full-segment spectral threshold is prone to false alarms under broadband impacts, and there is no segment-adaptive decomposition parameter design to balance steady-state and impact segments. Summary of the Invention

[0005] The purpose of this invention is to provide a method, system, and device for identifying faults in equipment comfort measurement, in order to solve one or more technical problems existing in the prior art, and at least provide a beneficial option or create conditions.

[0006] To achieve the above objectives, according to one aspect of the present invention, a method for identifying faults in equipment comfort measurement is provided, the method comprising the following steps:

[0007] The target sequence is obtained by acquiring continuous sampling signals of the device during one complete movement of the path to be detected;

[0008] Adaptive segment decoupling is performed on the target sequence to output multiple sampled signal segments and their time order;

[0009] For each sampled signal segment, variational mode decomposition yields multiple mode sequences with the same length and aligned position to the sampled signal segment.

[0010] For each sampled signal segment, the signal segment is traversed in position order to form a signal segment extension wave: at each position, the modal traversal prefix of each modal sequence up to that position is taken, the pairwise similarity between the modal traversal prefixes is calculated and the arithmetic mean is obtained to obtain the signal segment voltage value of that position; the signal segment voltage values ​​of each position are arranged in position order to form the signal segment extension wave of that sampled signal segment.

[0011] The peak values ​​of the extended wave of the signal segment are extracted, and the characteristics of the peak values ​​and their order are used to generate the extended wave ratio.

[0012] Record the original temporal order of each sampled signal segment; obtain the wave ratio order of each sampled signal segment from smallest to largest according to the extended wave ratio order; use the difference between the wave ratio order of each sampled signal segment and its original order as its wave ratio drop;

[0013] The signal segments with abnormal signal comfort were identified by filtering each segment based on the median wave ratio drop across all sampled signal segments.

[0014] The method described in this invention uses adaptive segment decoupling, including but not limited to the ASD algorithm, Jerk algorithm, and multi-model direct adaptive decoupling controller for stochastic systems. This is because in previous technologies, the boundary positions and durations of acceleration / deceleration and leveling fine-tuning of the same equipment under different loads and speed curves drifted with the operating conditions. Fixed thresholds or pure speed thresholds were often misclassified, easily leading to downstream diagnostic distortion. This invention is designed to extract core data features from quantile estimates of the rate of change of acceleration (jerk), hysteresis, and data such as the shortest duration / minimum gap. Robust statistics such as the multi-model direct adaptive decoupling controller for stochastic systems are then used to characterize the dynamic abrupt changes in stage transitions. Hysteresis avoids boundary jitter, and duration gap constraints retain short but realistic leveling segments, outputting each signal segment. This approach improves the boundary repeatability across loads and machines; for example, boundary jitter is significantly reduced, and the retention rate of short stages, especially leveling fine-tuning, is improved, reducing erroneous signal segments. Moreover, as an upstream module, it provides more homogeneous intra-segment statistics for subsequent VMD and feature extraction. The parameters in the VMD algorithm are set to be dynamically adaptive. For example, the number of modes K and bandwidth α are automatically adjusted according to the data of the section's operating conditions through various automatic algorithms. This is because previous VMD algorithms generally fixed the parameters K and α, which presented a dilemma in the steady-state and impact phases. A small K resulted in mixed modes, while a large K led to over-decomposition; an improperly preset α could cause noise and over-smoothing. Therefore, this method first estimates the number of spectral peaks and the signal-to-noise ratio, and then selects K and α accordingly, making the modal complexity of K and α approximately equal to the spectral complexity of the input data. For example, the bandwidth constraints of K and α are set to be inversely proportional to the noise level of the input data. This improves modal separability and the decomposition stability of different sections, thereby reducing reconstruction errors and modal center frequency drift.

[0015] Furthermore, the target sequence is obtained by processing the continuous sampling signals of the device during a complete movement of the path to be detected through attitude correction and second-order Butterworth filtering.

[0016] Furthermore, for each sampled signal segment, based on the number of spectral peaks and the signal-to-noise ratio within that sampled signal segment, the mode number and bandwidth constraint parameters in the variational mode decomposition algorithm are adaptively selected to obtain multiple mode sequences that are consistent with the length of each sampled signal segment and aligned in position.

[0017] Furthermore, in the process of forming the signal segment spread wave: the pairwise similarity adopts cosine similarity; when any mode ergodic prefix is ​​a zero sequence, the similarity with the other mode ergodic prefixes is treated as zero; the signal segment wave voltage value when the position is the first is set to zero; when multiple peak values ​​with the same value appear, the peak with the smaller position is selected to calculate the spread wave ratio and spread wave sequence.

[0018] Existing technologies, such as single-point peak values ​​and RMS, are point or full-segment statistics that cannot reflect the structural evolution within a segment, such as the accumulation or decay of impact modes with position sequence, and cannot be compared across modes. This invention, on position sequence t, takes the prefix of each mode, calculates pairwise similarity, and averages it to form X[t]; these are then sequentially arranged to obtain the extended wave Y. This construction quantifies the consistency and deviation on the position sequence into comparable structural curves and achieves cross-modal alignment. This approach is more sensitive to staged morphological changes, especially the start and end of impacts, and has better noise resistance and cross-modal consistency. Its similarity score remains high even when amplitudes differ but morphologies are similar, providing a stable structural basis for subsequent extended wave ratio μ, extended wave sequence no, and extended wave ratio sequence nummr.

[0019] Furthermore, for each sampled signal segment, variational mode decomposition (VMD) is used to obtain multiple mode sequences with the same length and aligned position to the sampled signal segment. The VMD algorithm incorporates the following parameters and constraints: when the number of spectral peaks in each sampled signal segment does not exceed one, the parameter representing the number of modes used in the VMD algorithm is set to two; when the number of spectral peaks is two to three, the number of modes is three to four; and when the number of spectral peaks is more than three, the number of modes is not less than four. For example, when the signal-to-noise ratio (SNR) is high, the bandwidth constraint parameter is set to the high range; when the SNR is moderate, it is set to the middle range; and when the SNR is low, it is set to the low range.

[0020] Further, the process of extracting peak values ​​and their positions from the spread wave of the signal segment to generate a spread wave ratio order specifically involves: performing peak detection on the spread wave of the signal segment; for each peak value, calculating the ratio between the peak value and the median of all values ​​in the spread wave of the signal segment as the spread wave ratio, and calculating the ratio between the position of the peak value and the total length of the spread wave of the signal segment as the spread wave order; using the product of the spread wave ratio and the spread wave order as the spread wave ratio order of the peak, and using the maximum value among the spread wave ratio orders of all peaks of each sampled signal segment as the spread wave ratio order of that segment.

[0021] Existing technologies exhibit incomparable peak amplitudes across different segment lengths and sampling densities, and the timing of peak appearance significantly impacts user experience. This invention designs the extended wave ratio μ, extended wave sequence no, and extended wave ratio sequence nummr to define μ using the relative amplitude of peak / median, avoiding incomparability in dimensions and amplitude. No is defined using the relative position of peak position / total length because later peak appearances may have a greater impact on the end-of-ride experience, among other engineering considerations. The product nummr then integrates the amplitude and positional data features. This approach facilitates comparability across different lengths and sampling rates while enhancing sensitivity to anomalies at the end of each sampled signal segment. As a sorting feature, it stably distinguishes between normal and slightly abnormal segments. Existing technologies typically only consider amplitude or energy, rarely incorporating peak position normalization into core scoring. This invention treats position normalization as an equally important factor, forming a structural strength indicator of amplitude fusion position, demonstrating clear application motivation and engineering significance.

[0022] Furthermore, the method may further include a component indication step: for each modal sequence identified as a signal comfort anomaly segment, data features such as its center frequency distribution, energy distribution, envelope impact characteristics, and harmonic sparsity are acquired as target features; the target features are mapped and matched with each component constituting the device using mapping algorithms, including but not limited to neural networks, to output the component category associated with the signal comfort anomaly segment and its confidence level. The component can be used to establish a correspondence between a specific frequency band or feature combination and a component category using an indication rule base, and threshold conditions or weighted scoring can be used as criteria. When the correspondence is satisfied, the corresponding component indication result can be output. When multiple mapped and matched components exist, each component is selected according to its confidence level.

[0023] In some embodiments, the system may also include a maintenance suggestion generation step: based on the component indication results and anomaly levels, generating corresponding maintenance suggestions and re-inspection plans from a maintenance suggestion rule base; the maintenance suggestion rule base maps component categories and anomaly levels to specific operation suggestions and time arrangements, which may include, but are not limited to: for example, for guide shoe related indications, suggestions for guide shoe clearance re-inspection, guide rail surface cleaning and lubrication, and fastener inspection and resetting; for guide rail joint related indications, suggestions for joint bolt tightening, end alignment verification, and contact surface repair; for traction machine related indications, suggestions for bearing lubrication status inspection, transmission component tension and wear verification, and speed-limited inspection when necessary; and providing preset short-term re-testing time windows and re-inspection data retention requirements in the suggestions.

[0024] Furthermore, the method may also include: recording abnormal segments of the same path or the same device during multiple runs; when abnormal signal segments pointing to the same component occur consecutively or multiple times, or when the cumulative number of occurrences within a preset period reaches a threshold, an instruction is given to automatically stop the machine for inspection or limit its speed.

[0025] This invention also provides a fault identification system for equipment comfort measurement. The system includes a processor, a memory, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps of the fault identification method for equipment comfort measurement. The system can run on computing devices such as desktop computers, laptops, handheld computers, and cloud data centers. The runnable system may include, but is not limited to, processors, memory, and server clusters. The processor executes the computer program within the following system units:

[0026] The modal serialization unit is used to acquire continuous sampled signals of the device during the entire movement process to obtain a target sequence; decouple the target sequence to output multiple sampled signal segments and their time order; for each sampled signal segment, multiple modal sequences with the same length and positional alignment as the sampled signal segment are obtained through variational mode decomposition;

[0027] The signal segment expansion unit is used to traverse each sampled signal segment in position order to form a signal segment expansion wave: at each position, the modal traversal prefix of each modal sequence up to that position is taken, the pairwise similarity between the modal traversal prefixes is calculated and the arithmetic mean is obtained to obtain the signal segment voltage value of that position; the signal segment voltage values ​​of each position are arranged in position order to form the signal segment expansion wave of that sampled signal segment;

[0028] The segment difference sorting unit is used to extract the peak value and its positional characteristics from the extended wave of the signal segment to generate the extended wave ratio order; record the original order of each sampled signal segment in time; obtain the wave ratio sort according to the extended wave ratio order of each sampled signal segment from smallest to largest; and use the difference between the wave ratio sort of each sampled signal segment and its original sort as its wave ratio drop.

[0029] The comfort anomaly segment identification unit is used to filter each sampled signal segment based on its wave ratio drop using the median of the wave ratio drop of all sampled signal segments, and identify signal comfort anomaly segments.

[0030] Correspondingly, the present invention also provides an electronic device, a readable storage medium, and a computer program product:

[0031] An electronic device includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the device comfort measurement fault identification method and the method for each step thereof.

[0032] A non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to cause the computer to perform the device comfort measurement fault identification method and the steps thereof.

[0033] A computer program product includes a computer program that, when executed by a processor, implements the device comfort measurement fault identification method and the methods for each step thereof.

[0034] The beneficial effects of this invention are as follows: This invention provides a method, system, and device for fault identification in equipment comfort measurement. It can collect continuous sampling signals such as triaxial velocity and angular velocity during a complete movement of the equipment along the path to be detected. Through adaptive segment decoupling, multiple sampling signal segments are obtained by merging constraints based on the quantile threshold of the rate of change of acceleration, hysteresis, shortest duration, and minimum gap. Variational mode decomposition (VMD) is performed on each sampling segment, and the number of modes and bandwidth constraints are adaptively selected based on the number of spectral peaks and signal-to-noise ratio of that segment, resulting in multiple mode sequences that are consistent with the segment length and aligned in position. The similarity of each mode prefix is ​​calculated pairwise according to position and averaged to form the signal segment wave volt X of that segment, which is arranged in position as the signal segment extended wave Y. Peak values ​​are extracted from the extended wave, and the extended wave ratio μ and extended wave sequence no are calculated. Their product is used to obtain the extended wave ratio sequence numr. Each sampling segment is sorted according to numr and compared with the original time sorting to obtain the wave ratio drop Lo. Abnormal segments are determined based on the γ-weighted threshold of the median and the median absolute deviation.

[0035] This invention is applicable to elevators, escalators, rail transit, and high-precision automatic driving motion platforms. It possesses advantages such as robustness across operating conditions, stable decomposition, sensitivity to structural evolution within segments, and interpretable maintenance implementation. For each sampling segment, the mode number and bandwidth constraints of VMD are adaptively set according to the number of spectral peaks and the signal-to-noise ratio, ensuring that modal complexity matches spectral complexity and bandwidth constraints self-adjust with noise. This alleviates the dilemma between broadband impacts and steady-state narrowband, reduces the risk of mixed modes and over-decomposition, improves the stability of modal center frequencies and energy characteristics, and facilitates comparability across repeated measurements. Based on the pairwise similarity of the modal prefixes of the signal segment wave volt X and extended wave Y, the positional structural changes that were previously difficult to express through point / full-segment statistics are transformed into alignable and comparable structural features. This enables sensitive capture of the start and end positions and intensity evolution of impacts, improving the detection capability of minor anomalies. The extended wave ratio (μ) is relativized in amplitude and the extended wave ratio sequence (no) is normalized in position to construct the extended wave ratio sequence (numr), thus achieving a structural strength measurement of amplitude and position. This measurement is insensitive to segment length, sampling rate, and amplitude scale, and is suitable for lateral comparison and ranking across different devices and sampling configurations. Structural strength is ranked using numr and compared with the original temporal ranking to obtain the wave ratio drop (Lo). A robust threshold is then formed using the median and median absolute deviation (γ-weighted). This decision mechanism has stronger fault tolerance for outliers and scene fluctuations, effectively suppressing false alarms and jitter alarms while maintaining the detection rate. Attached Figure Description

[0036] The above and other features of the present invention will become more apparent from the detailed description of the embodiments shown in conjunction with the accompanying drawings. In the accompanying drawings, the same reference numerals denote the same or similar elements. Obviously, the drawings described below are merely some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without any creative effort. In the drawings:

[0037] Figure 1 The diagram shows a flowchart of a method for identifying faults in equipment comfort measurement.

[0038] Figure 2 The figure shown is a system structure diagram of a fault identification system for measuring equipment comfort. Detailed Implementation

[0039] The following will provide a clear and complete description of the concept, specific structure, and technical effects of the present invention in conjunction with the embodiments and accompanying drawings, so as to fully understand the purpose, solution, and effects of the present invention. It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other.

[0040] In the description of this invention, "several" means one or more, "multiple" means two or more, "greater than," "less than," and "exceeding" are understood to exclude the stated number, while "above," "below," and "within" are understood to include the stated number. The use of "first" and "second" in the description is merely for distinguishing technical features and should not be construed as indicating or implying relative importance, or implicitly indicating the number of indicated technical features, or implicitly indicating the order of the indicated technical features.

[0041] like Figure 1 The diagram shown is a flowchart of a fault identification method for equipment comfort measurement according to the present invention. The following is a description of the method in conjunction with... Figure 1 This paper describes a method, system, and device for identifying equipment comfort measurement faults according to embodiments of the present invention.

[0042] This invention proposes a method for identifying faults in equipment comfort measurement, the method specifically including the following steps:

[0043] The target sequence is obtained by acquiring continuous sampling signals during the complete movement of the device; the target sequence is decoupled and output as multiple sampling signal segments and their time order; for each sampling signal segment, multiple mode sequences with the same length and positional alignment as the sampling signal segment are obtained by variational mode decomposition.

[0044] For each sampled signal segment, the signal segment is traversed in position order to form a signal segment extension wave: at each position, the modal traversal prefix of each modal sequence up to that position is taken, the pairwise similarity between the modal traversal prefixes is calculated and the arithmetic mean is obtained to obtain the signal segment voltage value of that position; the signal segment voltage values ​​of each position are arranged in position order to form the signal segment extension wave of that sampled signal segment.

[0045] The peak values ​​and their order are extracted from the extended wave of the signal segment to generate the extended wave ratio order; the original temporal order of each sampled signal segment is recorded; the wave ratio order is obtained according to the extended wave ratio order of each sampled signal segment from smallest to largest; the difference between the wave ratio order of each sampled signal segment and its original order is taken as its wave ratio drop.

[0046] The signal segments with abnormal signal comfort were identified by filtering each segment based on the median wave ratio drop across all sampled signal segments.

[0047] In some embodiments, taking an automatic elevator as an example, the device moves on the path to be detected, and during a complete movement on the path to be detected, a continuous sampling signal is obtained by continuously sampling the signals of the device, including its three-axis velocity and angular velocity.

[0048] The continuously sampled signal is divided into multiple different sampled signal segments through adaptive segment decoupling;

[0049] Using the VMD algorithm, each sampled signal segment is transformed into multiple modal sequences corresponding to each sampled signal segment;

[0050] In this context, the sequence lengths of the multiple modal sequences corresponding to the sampled signal segments are all consistent, and the values ​​of each bit sequence in each modal sequence are aligned with each other. The device refers to a mechanical device capable of performing at least one complete movement along a fixed path, and the device may include, but is not limited to, automatic elevators, automatic trams, escalators, automatic lifts, autonomous vehicles, autonomous drones, and automatic cable cars.

[0051] Furthermore, the pairwise similarity is cosine similarity; when any modal prefix is ​​a zero sequence, the similarity with the other modes is treated as zero; the signal segment volt value when the position is first is set to zero; when multiple peak values ​​with the same value appear, the peak with the smaller position is selected as the basis for calculating the spread wave ratio and spread wave sequence.

[0052] Furthermore, the number of modes and the bandwidth constraint parameter are determined according to the following rules: when the number of spectral peaks does not exceed one, the number of modes is two; when the number of spectral peaks is two to three, the number of modes is three to four; when the number of spectral peaks is more than three, the number of modes is not less than four; when the signal-to-noise ratio is high, the bandwidth constraint parameter is in the high range; when the signal-to-noise ratio is in the middle range, the bandwidth constraint parameter is in the middle range; and when the signal-to-noise ratio is low, the bandwidth constraint parameter is in the low range.

[0053] Example 1:

[0054] The bit order of each modal sequence corresponding to each sampled signal segment is traversed.

[0055] For each position in the current traversal,

[0056] Select the values ​​of each bit sequence from the first bit sequence to the currently traversed bit sequence in each modal sequence corresponding to the sampled signal segment.

[0057] An array consisting of the digits from the first position in each modal sequence to the current position is used as the prefix for each modal sequence corresponding to the current position. For example, if the current position is position 3, the three digits from the first to the third position in each modal sequence are used as the prefixes for each modal sequence corresponding to the third position in the current sequence.

[0058] Calculate the similarity between each pair of modal traversal prefixes corresponding to the current traversal position, and take the arithmetic mean of the calculated similarity values ​​as the signal segment volt(X) corresponding to the current traversal position of the sampled signal segment. However, the signal segment volt(X) value corresponding to the first position of each sampled signal segment is set to zero.

[0059] By traversing each bit sequence, the signal segment volt (X) corresponding to each bit sequence of the sampled signal segment is calculated and obtained in sequence. The signal segment volt (X) corresponding to each bit sequence of the sampled signal segment is arranged in order to form a sequence, which is used as the signal segment spread wave (Y) corresponding to the sampled signal segment.

[0060] A complete operation of the passenger elevator is performed with a fixed sampling rate; attitude correction and second-order Butterworth filtering are completed. Through adaptive segment decoupling, three sampled signal segments are obtained, denoted in chronological order as the first, second, and third sampling segments, with their original order being I, II, and III, respectively. Then, variational mode decomposition is performed independently on each sampling segment, resulting in three position-aligned mode sequences. The mode sequence length for each sampling segment is four sample positions.

[0061] Below are three simplified numerical examples of modal sequences, each with four positional values, where all units are dimensionless values ​​after normalization:

[0062] First sampling segment: Mode 1: [2, 2, 2, 2]; Mode 2: [2, 2, 2, 2]; Mode 3: [0, 1, 0, 1].

[0063] Second sampling segment: Mode 1: [2, 2, 2, 2]; Mode 2: [2, 2, 2, 2]; Mode 3: [1, 0, 1, 0].

[0064] The third sampling segment: Mode 1: [2, 2, 2, 2]; Mode 2: [2, 2, 2, 2]; Mode 3: [1, 1, 0, 0].

[0065] For each sampling segment, processing is performed sequentially from position one to position four. It is evident that the volt value for position one, etc., should be set to zero; however, starting from position two onwards, within the prefix interval preceding that position, the similarity of the prefix sequences of the three modes is calculated pairwise, and the arithmetic mean is taken as the volt value for that position. The volt values ​​of the four positions are then arranged in order to form the signal segment extension waveform.

[0066] First sampling segment: The extended wave is [0, 0.805, 0.718, 0.805].

[0067] Second sampling segment: The extended wave is [0, 0.805, 0.877, 0.805];

[0068] Third sampling segment: The extended wave is [0, 1.000, 0.877, 0.805].

[0069] For each sampling segment, peak detection is first performed on its spread wave. For each peak, the ratio between the peak value and the median of the spread wave is calculated as the spread wave ratio. The ratio between the position of the peak and the total length of the spread wave is used as the spread wave order. The spread wave ratio and the spread wave order are multiplied to obtain the spread wave ratio order of the peak. The maximum value among the spread wave ratio orders of all peaks in each sampling segment is taken as the representative value of the spread wave ratio order of this segment.

[0070] First sampling segment: The median of the spread wave is 0.7615; taking the peak value of 0.805 at position 2, the spread wave ratio is 1.057, the spread wave sequence is 0.50, and the representative value of the spread wave ratio sequence is 0.5285.

[0071] Second sampling segment: The median of the spread wave is 0.805; taking the peak value of 0.877 at position number three, the spread wave ratio is 1.0896, the spread wave sequence is 0.75, and the representative value of the spread wave ratio sequence is 0.8172.

[0072] Third sampling segment: The median of the spread wave is 0.841; taking the peak value of 1.000 at position 2, the spread wave ratio is 1.189, the spread wave sequence is 0.50, and the representative value of the spread wave ratio sequence is 0.5945.

[0073] Sort by extended wave ratio from smallest to largest, the wave ratios of the first, third, and second sampling segments are ordered as I, II, and III, respectively. Compared with the original order, the wave ratio drop of the first sampling segment is zero, the wave ratio drop of the third sampling segment is negative one, and the wave ratio drop of the second sampling segment is positive one.

[0074] The median of all wave ratio drops is used as the base threshold, and a weighted term corrected by the absolute deviation of the median is introduced, with the correction intensity coefficient denoted as γ. When the wave ratio drop of a certain sampling segment is greater than zero and greater than the corrected threshold, it is determined to be a signal abnormal segment. In this embodiment, when the median is zero and γ is 0.5, the second sampling segment meets the condition and is determined to be a signal abnormal segment.

[0075] This embodiment also provides the construction of two repeated tests: a normal group and a slightly abnormal group. The two groups are set up exactly the same, the only difference being a slight deviation in the gap between the guide shoe and the guide rail in the slightly abnormal group. Both groups are processed according to the same caliber as in Embodiment 1.

[0076] The comparative results show that in the normal group, the median of the spread spectrum in the first sampling segment is approximately 0.76. After selecting the peak value with the second position, the spread ratio is approximately 1.06, the spread spectrum order is 0.50, and the representative value of the spread ratio order is approximately 0.53. In the second sampling segment, the median of the spread spectrum is approximately 0.81. After selecting the peak value with the third position, the spread ratio is approximately 1.08, the spread spectrum order is 0.75, and the representative value of the spread ratio order is approximately 0.81. In the third sampling segment, the median of the spread spectrum is approximately 0.84. After selecting the peak value with the second position, the spread ratio is approximately 1.19, the spread spectrum order is 0.50, and the representative value of the spread ratio order is approximately 0.59. Correspondingly, the wave ratio difference in the normal group shows a pattern of zero in the first sampling segment, positive one in the second sampling segment, and negative one in the third sampling segment.

[0077] In the slightly anomalous group, the median of the spread wave in the first sampling segment was approximately 0.75. After selecting the peak with the second position, the spread wave ratio was approximately 1.08, the spread wave order was 0.50, and the representative value of the spread wave ratio order was approximately 0.54. In the second sampling segment, the median of the spread wave was approximately 0.78. After selecting the peak with the third position, the spread wave ratio increased to approximately 1.15, the spread wave order remained at 0.75, and the representative value of the spread wave ratio order increased to approximately 0.86. In the third sampling segment, the median of the spread wave was approximately 0.82. After selecting the peak with the second position, the spread wave ratio was approximately 1.20, the spread wave order was 0.50, and the representative value of the spread wave ratio order was approximately 0.60. Therefore, under the same threshold and γ value, the second sampling segment of the slightly anomalous group exhibited a higher spread wave ratio and spread wave ratio order compared to the normal group, resulting in a more significant shift in its order relative to the original sequence, and was stably identified as a signal anomalous segment. This is consistent with the micro-impact enhancement during the flat-layer stage.

[0078] For each sampling segment, the number of modes and bandwidth constraints are independently selected. Specifically, when the number of spectral peaks does not exceed one, the number of modes is two; when the number of spectral peaks is between two and three, the number of modes is three to four; and when the number of spectral peaks is more than three, the number of modes is not less than four. The bandwidth constraint parameter is inversely proportional to the noise level. Specifically, a higher bandwidth constraint parameter range is selected when the signal-to-noise ratio is high, a medium range is selected when the signal-to-noise ratio is moderate, and a lower range is selected when the signal-to-noise ratio is low, to enhance the separability between modes. This rule ensures that the length of each mode sequence within the same sampling segment is consistent and the position is aligned. The specific values ​​of the parameters can be automatically generated as a fixed configuration through calibration sets during the device deployment phase.

[0079] The γ coefficient used to weight the base threshold according to the median absolute deviation is recommended to be between zero and one. When the field noise is moderate, a value of 0.5 for γ can achieve a balance between sensitivity and robustness; when the field noise increases significantly or the number of sampling segments increases, γ can be appropriately increased to reduce the impact of occasional spikes on anomaly detection.

[0080] Example 2:

[0081] Align the signal segment spread waves (Y) corresponding to each sampled signal segment with each other, and ensure that the sequence lengths of the signal segment spread waves (Y) corresponding to each sampled signal segment are consistent.

[0082] For each signal segment spread wave (Y), the ratio of each peak value to the median or arithmetic mean of the values ​​of each position in the signal segment spread wave (Y) is obtained as the spread wave ratio (mu). The position number of each spread wave ratio (mu) in its respective signal segment spread wave (Y) is recorded. The ratio of the position number of each spread wave ratio (mu) to the length of the signal segment spread wave (Y) sequence is obtained as the spread wave order (no). The product of each spread wave ratio (mu) and its corresponding spread wave order (no) is calculated to generate the spread wave ratio order (numr) corresponding to that sampled signal segment.

[0083] Example 3:

[0084] First, record the original sequence number of each sampled signal segment as the original sorting of each sampled signal segment;

[0085] Then, the sampled signal segments are sorted from smallest to largest according to the extended wave ratio (numr) values, resulting in the wave ratio sort (ru sort) for each sampled signal segment.

[0086] Subtract the original sort from the wave ratio sort (ru sort) corresponding to each sampled signal segment to obtain the wave ratio drop (Lo) corresponding to each sampled signal segment.

[0087] In fact, existing technologies suffer from significant differences in numerical scales across different devices and rounds, making fixed thresholds prone to false alarms; simultaneously, it is necessary to consider abnormal displacements relative to the original time sequence. This invention first uses `numr` to sort the structural strength, then compares it with the original time sequence to obtain the drop (Lo), and uses the median, or with the added median absolute deviation (e.g., a weighting coefficient γ), to construct a robust threshold. Traditional alarms directly use thresholds or decouple statistics from the original sequence, while this method uses the sorting difference as an anomaly quantification and then uses a robust statistical threshold for cross-scenario application. The method described in this invention possesses adaptability and robustness across multiple devices and rounds, intuitively identifying how much the abnormal segment has been pushed back in the sorting.

[0088] Example 4:

[0089] The median of the wave ratio drop (Lo) for each sampled signal segment is obtained. Sampling signal segments with a wave ratio drop (Lo) greater than zero and greater than the median wave ratio drop (Lo) are selected as signal comfort anomaly segments. In some embodiments, the median wave ratio drop of all sampled signal segments can be used as a base threshold, and a weighted correction is applied based on the median absolute deviation, with a correction intensity coefficient of γ. When the wave ratio drop of a certain sampled signal segment is greater than zero and greater than the corrected threshold, that sampled signal segment is determined to be a signal comfort anomaly segment. The signal comfort anomaly segment can be used to represent sampling signal segments identified based on the wave ratio ranking and wave ratio drop of the sampled signal segments, where the user comfort of the device's operating angular velocity, angular velocity, etc., may be abnormal.

[0090] Furthermore, the method may further include a component indication step: for each modal sequence identified as an abnormal signal segment, obtain target features such as center frequency distribution, energy distribution, envelope impulse characteristics, and harmonic sparsity; match the target features with a preset component indication rule library, and output the component category associated with the abnormality and its confidence level, wherein the component indication rule library establishes a correspondence between specific frequency bands or feature combinations and component categories, and uses threshold conditions or weighted scoring as criteria, and outputs the corresponding component indication result when the correspondence is satisfied; when there are multiple candidate components, the priority indication result is output according to the higher confidence level or score.

[0091] Furthermore, the process may also include a maintenance suggestion generation step: based on the component indication results and anomaly levels, generating corresponding maintenance suggestions and re-inspection plans from the maintenance suggestion rule base; the maintenance suggestion rule base maps component categories and anomaly levels to specific operation suggestions and time arrangements, including but not limited to: for guide shoe related indications, providing suggestions for guide shoe clearance re-inspection, guide rail surface cleaning and lubrication, and fastener inspection and resetting; for guide rail joint related indications, providing suggestions for joint bolt tightening, end alignment verification, and contact surface repair; for traction machine related indications, providing suggestions for bearing lubrication status inspection, transmission component tension and wear verification, and speed-limited inspection when necessary; and providing preset short-term re-testing time windows and re-inspection data retention requirements in the suggestions.

[0092] Furthermore, abnormal segments of the same path or the same equipment during multiple runs are recorded longitudinally. When abnormal segments pointing to the same component category occur repeatedly or the cumulative number of occurrences reaches a threshold within a preset period, the maintenance recommendation is automatically upgraded to enhanced re-inspection, shutdown inspection, or speed-limited operation, and an abnormal work order and responsibility notification are triggered.

[0093] In some embodiments, a workflow for mapping signal characteristics of an abnormal segment to various components may also be provided. Specifically, the input consists of target data and modal characteristics of the segment identified as an abnormal signal segment, which may include at least: the center frequency distribution and energy distribution of each mode; the envelope impact characteristics of each mode or a specified frequency band; harmonic sparsity, spectral centroid, and whether there is an ordered harmonic overtone structure; segment operation status labels, such as acceleration, constant speed, deceleration, and leveling fine-tuning; and comparative statistics with the historical data of the same machine for consistency comparison to improve confidence.

[0094] The processing flow specifically includes: scaling the center frequency, energy ratio, envelope impact significance, and harmonic sparsity to obtain high, medium, and low level or abnormal, suspicious, and normal numerical indicators. Matching is performed one by one according to the corresponding mapping relationship between feature patterns and component categories; when multiple rules are satisfied simultaneously, a weighted scoring and priority strategy is used for aggregation. When the numerical indicator scores of mappings for two or more component categories are close, discrimination is made based on the operating status label and position; for example, if the abnormal segment mainly occurs during leveling fine-tuning and is characterized by a mid-to-high frequency envelope impact increase while low-frequency energy changes are not significant, then guide shoes and guide rail joints are given priority; if the abnormal segment mainly occurs during acceleration and deceleration and low-to-medium frequency energy fluctuates significantly with the load, then the candidate priority of the traction machine and drive chain is increased. Then, the confidence levels of high, medium, and low are given based on the ranking of the comprehensive numerical indicators. Finally, the component category, confidence level, main triggering characteristics, and visual prompts obtained from the mapping are output, such as sending a message to suggest viewing the envelope curve of the mapped frequency band.

[0095] For example, during the leveling fine-tuning phase, a concentrated mid-to-high frequency envelope impact rise occurs, with an irregular harmonic structure, high harmonic sparsity, and a center frequency distribution that remains similar with subsequent retests. This can be matched to indicate a probability of guide shoe clearance deviation or insufficient lubrication. For example, near the leveling position, the envelope energy proportion of a specified frequency band is significantly higher than the mid-level of the machine, exhibiting repetitive impacts with stable positional order; the rule base can point to a probability of insufficient joint tightening or uneven end faces. For example, during acceleration or deceleration, low to mid-frequency energy shows increased correlation with load and speed changes, while an ordered harmonic structure appears; thus, a command can be sent indicating a probability of traction machine bearing lubrication deterioration or uneven tension. For example, during the constant speed phase, slow, fluctuating low-frequency energy fluctuations are superimposed with intermittent subharmonics, and are correlated with load distribution; this can be used to send a command indicating a probability of rope groove wear or compensating chain oscillation.

[0096] In some embodiments, initial thresholds and priorities are generated automatically using machine learning algorithms during initial deployment, and continuously iterated based on re-inspection results, replacement records, and manual confirmation throughout the maintenance cycle. Version updates are implemented in a gray-scale manner, allowing for the retention of previous versions for retrospective analysis and comparison. Sometimes, the system also saves a reference index for each component indication result, facilitating future audits and technical comparisons.

[0097] The aforementioned equipment comfort measurement fault identification system operates in any computing device, such as a desktop computer, laptop computer, handheld computer, or cloud data center. The computing device includes a processor, a memory, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements the steps in the aforementioned equipment comfort measurement fault identification method. The operable system may include, but is not limited to, a processor, a memory, and a server cluster.

[0098] An embodiment of the present invention provides a fault identification system for measuring equipment comfort, such as... Figure 2 As shown, an embodiment of a device comfort measurement fault identification system includes: a processor, a memory, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps described in the embodiment of the device comfort measurement fault identification method. The processor executes the computer program within the following system unit:

[0099] The modal serialization unit is used to acquire continuous sampled signals of the device during the entire movement process to obtain a target sequence; decouple the target sequence to output multiple sampled signal segments and their time order; for each sampled signal segment, multiple modal sequences with the same length and positional alignment as the sampled signal segment are obtained through variational mode decomposition;

[0100] The signal segment expansion unit is used to traverse each sampled signal segment in position order to form a signal segment expansion wave: at each position, the modal traversal prefix of each modal sequence up to that position is taken, the pairwise similarity between the modal traversal prefixes is calculated and the arithmetic mean is obtained to obtain the signal segment voltage value of that position; the signal segment voltage values ​​of each position are arranged in position order to form the signal segment expansion wave of that sampled signal segment;

[0101] The segment difference sorting unit is used to extract the peak value and its positional characteristics from the extended wave of the signal segment to generate the extended wave ratio order; record the original order of each sampled signal segment in time; obtain the wave ratio sort according to the extended wave ratio order of each sampled signal segment from smallest to largest; and use the difference between the wave ratio sort of each sampled signal segment and its original sort as its wave ratio drop.

[0102] The comfort anomaly segment identification unit is used to filter each sampled signal segment based on its wave ratio drop using the median of the wave ratio drop of all sampled signal segments, and identify signal comfort anomaly segments.

[0103] In order to better unify the linear relationship and probabilistic connection between physical quantities with different units of measurement, dimensionless processing can be performed on different physical quantities.

[0104] Preferably, all undefined variables in this invention, if not explicitly defined, can be manually set thresholds.

[0105] The aforementioned equipment comfort measurement fault identification system can run on computing devices such as desktop computers, laptops, handheld computers, and cloud data centers. This system includes, but is not limited to, a processor and memory. Those skilled in the art will understand that the examples described are merely illustrations of a method, system, and device for identifying equipment comfort measurement faults, and do not constitute a limitation on such a method, system, and device. The system may include more or fewer components, combinations of certain components, or different components. For example, the system may also include input / output devices, network access devices, buses, etc.

[0106] The present invention also provides an electronic device, a readable storage medium, and a computer program product:

[0107] An electronic device includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the device comfort measurement fault identification method and the method for each step thereof.

[0108] A non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to cause the computer to perform the device comfort measurement fault identification method and the steps thereof.

[0109] A computer program product includes a computer program that, when executed by a processor, implements the device comfort measurement fault identification method and the methods for each step thereof.

[0110] The term "electronic device" is intended to refer to various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. Electronic devices can also refer to various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.

[0111] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0112] The program code used to implement the methods of the present invention can be written in any combination of one or more programming languages. This program code can be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing device, such that when executed by the processor or controller, the program code causes the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code can be executed entirely on the machine, partially on the machine, as a standalone software package partially on the machine and partially on a remote machine, or entirely on a remote machine or server.

[0113] In the context of this invention, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. Machine-readable media can include, but are not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0114] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device for displaying information to the user (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor); and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the computer. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0115] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as a data server), or computing systems that include middleware components (e.g., an application server), or computing systems that include frontend components (e.g., a user computer with a graphical user interface or web browser through which a user can interact with embodiments of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., a communication network). Examples of communication networks include local area networks (LANs), wide area networks (WANs), and the Internet.

[0116] Computer systems can include clients and servers. Clients and servers are generally located far apart and typically interact through communication networks. Client-server relationships are created by computer programs running on the respective computers and having a client-server relationship with each other.

[0117] The processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete component gate circuits, transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor. This processor is the control center of the equipment comfort measurement fault identification system, connecting various sub-regions of the system via various interfaces and lines.

[0118] The memory can be used to store the computer programs and / or modules. The processor, by running or executing the computer programs and / or modules stored in the memory and calling the data stored in the memory, realizes various functions of the device comfort measurement fault identification method, system, and device. The memory may mainly include a program storage area and a data storage area. The program storage area may store the operating system, at least one application program required for a function (such as sound playback function, image playback function, etc.), etc.; the data storage area may store data created based on the use of the mobile phone (such as audio data, phonebook, etc.). In addition, the memory may include high-speed random access memory, and may also include non-volatile memory, such as hard disk, memory, plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, at least one disk storage device, flash memory device, or other volatile solid-state storage device.

[0119] It should be understood that the various forms of processes shown above can be used to rearrange, add, or delete steps. For example, the steps described in this disclosure can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution disclosed in this invention can be achieved, and this is not limited herein.

[0120] This invention provides a method, system, and device for identifying faults in equipment comfort measurement. The method involves acquiring continuous sampled signals from the equipment during its complete movement to obtain a target sequence; decoupling the target sequence to output multiple sampled signal segments and their time order; using variational mode decomposition to obtain multiple mode sequences with the same length and aligned position of the sampled signal segments; traversing each sampled signal segment according to its position order to form a signal segment extension wave; extracting the peak values ​​and their positional characteristics from the signal segment extension wave to generate an extension wave ratio order; using the difference between the wave ratio order of each sampled signal segment and its original order as its wave ratio drop; and filtering each sampled signal segment based on its wave ratio drop using the median of the wave ratio drops of all sampled signal segments to identify signal comfort anomalies. This method is not sensitive to amplitude fluctuations in random noise but is more sensitive to envelope impacts and positional shifts in specific frequency bands, allowing for early detection of minor problems such as guide shoe clearance, guide rail joint impacts, and traction machine transmission abnormalities, facilitating preventative maintenance.

[0121] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A method for identifying faults in equipment comfort measurement, characterized in that, The method includes: The target sequence is obtained by acquiring continuous sampling signals of the device during its complete movement along the path to be detected; Adaptive segment decoupling is performed on the target sequence to output multiple sampled signal segments and their time order; For each sampled signal segment, variational mode decomposition yields multiple mode sequences with the same length and aligned position to the sampled signal segment. For each sampled signal segment, the signal segment is traversed in position order to form a signal segment extension wave. Specifically, at each position, the modal sequence up to that position is taken as the modal traversal prefix. The pairwise similarity between the modal traversal prefixes is calculated and averaged to obtain the signal segment voltage value at that position. The signal segment voltage values ​​at each position are arranged in position order to form the signal segment extension wave of the sampled signal segment. The characteristics of the peak values ​​and their order are extracted from the extended wave of the signal segment to generate the extended wave ratio; Record the original order of each sampled signal segment in time sequence; obtain the wave ratio sorting according to the extended wave ratio order of each sampled signal segment from smallest to largest; use the difference between the wave ratio sorting of each sampled signal segment and its original sorting as its wave ratio drop; The signal segments with abnormal signal comfort were identified by filtering each segment based on the median wave ratio drop of all sampled signal segments. Specifically, the process of extracting the peak values ​​and their order from the extended wave of the signal segment to generate an extended wave ratio is as follows: Peak detection is performed on the spread wave of the signal segment; for each peak, the ratio between the peak value and the median of all values ​​of the spread wave of the signal segment is calculated as the spread wave ratio, and the ratio between the position of the peak value and the total length of the spread wave of the signal segment is calculated as the spread wave sequence. The product of the spread wave ratio and the spread wave sequence is used as the spread wave ratio sequence of the peak, and the maximum value among the spread wave ratio sequences of all peaks of each sampled signal segment is used as the spread wave ratio sequence of that segment.

2. The method for fault identification in equipment comfort measurement according to claim 1, characterized in that, in, The target sequence is obtained by processing the continuous sampling signals of the device during a complete movement of the path to be detected through attitude correction and second-order Butterworth filtering.

3. The method for fault identification in equipment comfort measurement according to claim 2, characterized in that, in, For each sampled signal segment, based on the number of spectral peaks and the signal-to-noise ratio within that segment, the variational mode decomposition algorithm adaptively selects the mode number and bandwidth constraint parameters to obtain multiple mode sequences that are consistent with the length of each sampled signal segment and aligned in position.

4. The method for fault identification in equipment comfort measurement according to claim 1, characterized in that, in, During the formation of the signal segment spread wave: The pairwise similarity is cosine similarity; when any modal ergodic prefix is ​​a zero sequence, the similarity with the other modal ergodic prefixes is treated as zero; the signal segment volt value when the position is first is set to zero; when multiple peak values ​​with the same value appear, the peak with the smaller position is selected to calculate the spread wave ratio and spread wave sequence.

5. The method for fault identification in equipment comfort measurement according to claim 1, characterized in that, in, For each sampled signal segment, variational mode decomposition yields multiple mode sequences of the same length and aligned in position to the sampled signal segment. The variational mode decomposition algorithm also incorporates the following parameters and constraints: When the number of spectral peaks is no more than one, the number of modes is two; when the number of spectral peaks is two to three, the number of modes is three to four; when the number of spectral peaks is more than three, the number of modes is no less than four.

6. A method for identifying equipment comfort measurement faults according to any one of claims 1, 4, or 5, characterized in that, The method further includes the following steps: for each modal sequence identified as a signal comfort anomaly segment, feature extraction is performed to obtain its target features; The target feature is mapped and matched with each component constituting the device, and the component category and its confidence level associated with the signal comfort anomaly segment are output; when there are multiple mapped and matched components, each component is selected according to its confidence level.

7. The method for fault identification in equipment comfort measurement according to claim 1, characterized in that, in, Record abnormal segments of the same path or the same device during multiple runs. When abnormal signal segments pointing to the same component occur consecutively or multiple times, or when the cumulative number of occurrences within a preset period reaches a threshold, instruct the device to automatically stop for inspection or operate at a limited speed.

8. A fault identification system for measuring equipment comfort, characterized in that, The device comfort measurement fault identification system operates in any computing device, such as a desktop computer, a laptop computer, or a cloud data center. The computing device includes a processor, a memory, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements the steps of the device comfort measurement fault identification method as described in any one of claims 1 to 5.

9. An electronic device, comprising: At least one processor; and a memory communicatively connected to the at least one processor; The memory stores instructions executable by the at least one processor, characterized in that the instructions are executed by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 5.

Images