Vertical shaft rigid shaft guide fault diagnosis method and system based on gyroscope

Abstract

The invention relates to the technical field of fault diagnosis, in particular to a vertical shaft rigid cage guide fault diagnosis method and system based on a gyroscope, and the method comprises the steps: firstly, carrying out the preprocessing of original angular velocity signals of an X axis, a Y axis and a Z axis, and obtaining the corresponding preprocessed signals of each axis; secondly, parallel double-path analysis is carried out on the preprocessed signals corresponding to all the axes, and frequency domain features and time-frequency domain features corresponding to all the axes are obtained; splicing the frequency domain features and the time-frequency domain features corresponding to the three axes according to a predetermined sequence to obtain a comprehensive fault feature vector; and finally, the fault type of the rigid cage guide is diagnosed according to the comprehensive fault feature vector, the main direction of abnormal angular motion caused by the fault is judged, and a diagnosis result is output. According to the method, FFT and CWT are adopted for parallel double-path analysis, more comprehensive fault features can be extracted, and therefore the identification accuracy of the two fault types of joint dislocation and surface protrusion is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of fault diagnosis, and in particular to a gyroscope-based vertical shaft rigid tankway fault diagnosis method and system. Background Art

[0002] Hoisting systems are essential, critical equipment in mining, construction, and other fields. Their safe and stable operation is crucial for ensuring production and personnel safety. The condition of the hoistway, the guiding component of the hoist vessel, directly impacts the stability and safety of the hoist system. During operation, various factors can cause defects such as misaligned joints, protrusions, gaps, or wavy wear on the hoistway surface. If these defects are not discovered and addressed promptly, they can cause severe vibration of the hoist vessel, hinder operation, and even lead to serious safety accidents.

[0003] Current methods for monitoring tankway status primarily include manual inspections, static geometric measurements, and sensor-based dynamic monitoring. Manual inspections are inefficient, subjective, and difficult to achieve continuous monitoring. While static geometric measurements can obtain the geometric parameters of the tankway, they cannot reflect its dynamic characteristics under high-speed operation. Sensor-based dynamic monitoring methods have been widely used in recent years. They utilize sensors for data acquisition and posture estimation, with the goal of indirectly inferring tankway deformation by monitoring changes in the posture of the lifting container. However, these sensor-based dynamic monitoring methods, which indirectly infer tankway deformation, suffer from low monitoring accuracy.

[0004] While traditional vibration analysis methods are mature, they still present challenges when applied directly to data from rotational sensors such as gyroscopes to specifically diagnose tankway joint misalignment and surface defects. Urgent challenges lie in effectively isolating and identifying the subtle signatures of these specific defects from gyroscope signals, which contain multiple motion components and noise. Furthermore, how can these signatures accurately determine the fault type and determine the primary direction of abnormal rotation caused by the fault? Summary of the Invention

[0005] The purpose of the present invention is to provide a gyroscope-based method and system for diagnosing faults in a vertical shaft rigid tankway, so as to accurately diagnose the type of fault and the direction of abnormal angular motion caused by the fault.

[0006] To achieve the above object, the present invention provides a gyroscope-based method for diagnosing faults in a vertical shaft rigid tankway, the method comprising:

[0007] Get the original angular velocity signals of X, Y and Z axes;

[0008] Preprocessing the original angular velocity signal of each axis respectively to obtain a preprocessed signal corresponding to each axis;

[0009] Performing parallel dual-path analysis on the pre-processed signals corresponding to each axis to obtain frequency domain features and time-frequency domain features corresponding to each axis;

[0010] The frequency domain features and time-frequency domain features corresponding to the three axes are spliced ​​in a predetermined order to obtain a comprehensive fault feature vector;

[0011] The fault type of the rigid tank channel is diagnosed according to the comprehensive fault characteristic vector, and the main direction of the abnormal angular motion caused by the fault is determined, and the diagnosis result is output.

[0012] Optionally, preprocessing the original angular velocity signal of each axis to obtain a preprocessed signal corresponding to each axis specifically includes:

[0013] Daubechies 4 wavelet is selected as the mother wavelet, and the original angular velocity signal x of each axis is raw,c (t) Perform L-layer wavelet decomposition to obtain the detail coefficients cD of each layer l and the approximation coefficient cA L ;

[0014] Calculate the noise standard deviation based on the median estimation method;

[0015] Determine the threshold corresponding to the detail coefficient of each layer according to the noise standard deviation;

[0016] A soft threshold function with a threshold corresponding to each layer detail coefficient is used to perform threshold shrinkage processing on each layer detail coefficient to obtain the detail coefficient cD′ after threshold shrinkage processing. l ;

[0017] The detail coefficients cD′ of each layer after the threshold shrinkage process l and the approximation coefficient cA L Perform wavelet reconstruction to obtain the preprocessed signal x c (t).

[0018] Optionally, the specific formula of the soft threshold function is:

[0019] cD′ l =sgn(cD l )·max(0,|cD l |-δ l )

[0020] Among them, cD′ l is the detail coefficient of the lth layer after threshold shrinkage processing, δ l is the threshold corresponding to the detail coefficient of the lth layer, cD l is the detail coefficient corresponding to the lth layer, sgn() is the sign function, max() is the maximum value function, and 0 is to ensure that the minimum value of the coefficient is not negative.

[0021] Optionally, performing parallel dual-path analysis on the preprocessed signals corresponding to the axes to obtain frequency domain features and time-frequency domain features corresponding to the axes specifically includes:

[0022] Frequency domain analysis of the first path: The Welch method is used to calculate the power spectral density of the preprocessed signal, and a power spectral density graph of each axis is plotted based on the power spectral density. A peak detection algorithm is then used to extract n frequency peaks with maximum energy and their corresponding amplitudes from the power spectral density graph as frequency domain features, i.e., FFT features. Here, n is a positive integer greater than 1.

[0023] Time-frequency domain analysis of the second path: complex Morlay wavelet is used as the mother wavelet, and continuous wavelet transform (CWT) processing is performed on the preprocessed signal to obtain time-frequency energy. A time-frequency energy graph is drawn based on the time-frequency energy, and local energy maxima are searched on the time-frequency energy graph to identify m transient impact events. The occurrence time, center frequency and peak energy corresponding to each transient impact event are used as time-frequency domain features, namely CWT features; where m is a positive integer greater than 1.

[0024] Optionally, the frequency domain features and time-frequency domain features corresponding to the three axes are spliced ​​in a predetermined order to obtain a comprehensive fault feature vector. The specific formula is:

[0025]

[0026] Among them, f F,c,i ,A F,c,i The frequency peak value and amplitude corresponding to the peak value of the i-th peak energy on the c-th axis are sorted from high to low according to the peak energy, E C,c,j ,f C,c,j ,t C,c,j are the peak energy, center frequency and occurrence time corresponding to the jth transient impact time on the cth axis, respectively. c∈{X,Y,Z}, where X,Y,Z are the three axes.

[0027] Optionally, the diagnosing the fault type of the rigid tank channel according to the comprehensive fault feature vector specifically includes:

[0028] The fault type of the rigid tank channel is diagnosed by using the constructed comprehensive feature vector and a preset classification logic;

[0029] Alternatively, the fault type of the rigid tank channel may be diagnosed by using the constructed comprehensive feature vector and a trained machine learning model.

[0030] The present invention also provides a gyroscope-based vertical shaft rigid tankway fault diagnosis system, the system comprising:

[0031] The acquisition module is used to obtain the original angular velocity signals of the X, Y, and Z axes;

[0032] A preprocessing module, configured to preprocess the original angular velocity signal of each axis to obtain a preprocessed signal corresponding to each axis;

[0033] A dual-path analysis module is used to perform parallel dual-path analysis on the pre-processed signals corresponding to each axis to obtain frequency domain features and time-frequency domain features corresponding to each axis;

[0034] A splicing module is used to splice the frequency domain features and time-frequency domain features corresponding to the three axes in a predetermined order to obtain a comprehensive fault feature vector;

[0035] The fault diagnosis module is used to diagnose the fault type of the rigid tank channel according to the comprehensive fault characteristic vector, determine the main direction of abnormal angular movement caused by the fault, and output the diagnosis result.

[0036] Optionally, the dual-path analysis module includes:

[0037] The frequency domain analysis unit of the first path is used to calculate the power spectral density of the preprocessed signal using the Welch method, draw a power spectral density graph of each axis based on the power spectral density, and extract n frequency peaks with maximum energy and the amplitudes corresponding to the frequency peaks as frequency domain features, i.e., FFT features, from the power spectral density graph using a peak detection algorithm; wherein n is a positive integer greater than 1;

[0038] The time-frequency domain analysis unit of the second path is used to use the complex Morlay wavelet as the mother wavelet, perform continuous wavelet transform (CWT) processing on the preprocessed signal to obtain time-frequency energy, draw a time-frequency energy graph based on the time-frequency energy, search for local energy maxima on the time-frequency energy graph to identify m transient impact events, and use the occurrence time, center frequency and peak energy corresponding to each transient impact event as time-frequency domain features, i.e., CWT features; wherein m is a positive integer greater than 1.

[0039] The present invention also provides an electronic device, comprising a storage medium, a processor, and a computer program stored on the storage medium and executable on the processor, wherein the processor implements the above-mentioned method when executing the computer program.

[0040] The present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program implements the above method when executed by a processor.

[0041] According to the specific embodiments provided by the present invention, the present invention discloses the following technical effects:

[0042] The present invention provides a method and system for diagnosing faults in a vertical shaft rigid tankway based on a gyroscope. First, the original angular velocity signals of the X, Y, and Z axes are preprocessed separately to obtain the preprocessed signals corresponding to each axis. Second, the preprocessed signals corresponding to each axis are subjected to parallel dual-path analysis to obtain the frequency domain features and time-frequency domain features corresponding to each axis. The frequency domain features and time-frequency domain features corresponding to the three axes are then spliced ​​in a predetermined order to obtain a comprehensive fault feature vector. Finally, the fault type of the rigid tankway is diagnosed based on the comprehensive fault feature vector, and the main direction of abnormal angular motion caused by the fault is determined, and the diagnosis result is output. The present invention uses FFT and CWT for parallel dual-path analysis, which can extract more comprehensive fault features, thereby improving the recognition accuracy of the two fault types of joint dislocation and surface protrusion. The present invention can evaluate the main direction of the fault impact or the direction of abnormal change in the tank cage posture by comprehensively analyzing the response characteristics of the three-axis gyroscope signal. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0044] Figure 1 This is a flow chart of a method for diagnosing faults in a vertical shaft rigid tankway based on a gyroscope according to an embodiment of the present invention;

[0045] Figure 2 This is a structural diagram of the vertical shaft rigid tankway fault diagnosis process based on a gyroscope according to an embodiment of the present invention. DETAILED DESCRIPTION

[0046] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0047] The purpose of the present invention is to provide a gyroscope-based method and system for diagnosing faults in a vertical shaft rigid tankway, so as to accurately diagnose the type of fault and the direction of abnormal angular motion caused by the fault.

[0048] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0049] like Figure 1-2As shown, the present invention provides a gyroscope-based vertical shaft rigid tankway fault diagnosis method, the method comprising:

[0050] Step S1: Obtain the original angular velocity signals of the X, Y and Z axes.

[0051] Step S2: pre-processing the original angular velocity signal of each axis respectively to obtain a pre-processed signal corresponding to each axis.

[0052] Step S3: performing parallel dual-path analysis on the pre-processed signals corresponding to each axis to obtain frequency domain features and time-frequency domain features corresponding to each axis.

[0053] Step S4: The frequency domain features and time-frequency domain features corresponding to the three axes are spliced ​​in a predetermined order to obtain a comprehensive fault feature vector.

[0054] Step S5: diagnose the fault type of the rigid tank channel according to the comprehensive fault feature vector, determine the main direction of abnormal angular motion caused by the fault, and output the diagnosis result.

[0055] The following describes each step in detail:

[0056] Step S1: Get the original angular velocity signals of the X, Y and Z axes, denoted as x raw,c (t), where c∈{X, Y, Z}; the present invention discloses using an MPU6050 inertial measurement unit to obtain original angular velocity signals of the X, Y, and Z axes, wherein the MPU6050 inertial measurement unit is installed on the lifting container and the sampling frequency is set to 1 kHz.

[0057] The present invention also includes pre-filtering the original angular velocity signals of each axis. The purpose of pre-filtering is to suppress the sensor's own noise, electromagnetic interference in the operating environment, and high-frequency vibration components unrelated to the tank fault. At the same time, it is necessary to retain the useful signal characteristics related to the target fault, especially the transient impact caused by surface protrusions, to the greatest extent possible, especially the high-frequency components and mutation details that are crucial for subsequent transient feature analysis.

[0058] Furthermore, according to the method of the present invention, pre-filtering may be performed by one or a combination of the following methods:

[0059] Based on first-order or second-order complementary filters, the filter parameters such as time constant or cutoff frequency need to be finely optimized to preserve fault characteristics, especially high-frequency transient characteristics.

[0060] An adaptive Kalman filter is used, which can dynamically adjust its internal parameters according to the real-time statistical characteristics of the input signal, such as changes in the noise level, thereby achieving better noise suppression effects in complex and time-varying mine environments.

[0061] The multi-resolution analysis characteristics of the wavelet transform are used to decompose and reconstruct the signal at different scales. By thresholding the wavelet coefficients at different scales, noise can be effectively removed while preserving the non-stationary characteristics and transient impact components in the signal. This is crucial for the subsequent accurate extraction of fault features such as surface protrusions.

[0062] In order to effectively suppress the high-frequency random noise generated by electromagnetic interference while maximally retaining the transient characteristics generated by the fault impact, the present invention preferably uses a wavelet denoising method to denoise the three-axis original angular velocity signals.

[0063] Step S2: Preprocessing the original angular velocity signal of each axis to obtain the preprocessed signal corresponding to each axis, specifically including:

[0064] Step S21: select Daubechies 4 wavelet as the mother wavelet, and transform the original angular velocity signal x of each axis into raw,c (t) Perform L-layer wavelet decomposition to obtain the detail coefficients of each layer and the approximation coefficient cA L Where L can be a positive integer greater than 1; cD l is the detail coefficient corresponding to the lth layer, and the value range of l is 1 to L. In this embodiment, 5 layers are preferred.

[0065] The Daubechies 4 wavelet is chosen as the mother wavelet in this paper because it performs well on impulse signals. In the field of signal processing, the wavelet transform is a commonly used time-frequency analysis method. When a signal is decomposed using wavelet transforms, it is broken down into components of different frequency bands. The approximation coefficients represent the low-frequency portion of the signal, reflecting its primary characteristics; the detail coefficients represent the high-frequency portion of the signal, primarily reflecting its detailed information.

[0066] Step S22: Calculate the noise standard deviation using the median estimation method. The specific formula is: Among them, δ c is the noise standard deviation, median() is the median operation, W d1c is the wavelet decomposition coefficient.

[0067] Step S23: According to the noise standard deviation δ c Determine the threshold δ corresponding to each layer detail coefficient l , the specific formula is: in, Indicates the parameter δ that takes the minimum value of the expression in the brackets and δ is not a negative number. δ represents the candidate threshold tested during the optimization process. N represents the total number of wavelet detail coefficients in the i-th layer currently being processed. iIndicates the value of the i-th wavelet coefficient in the current layer, Count() is the counting function, and min() is the minimum value function.

[0068] Step S24: Using a soft threshold function with a threshold corresponding to each layer detail coefficient, perform threshold shrinkage processing on each layer detail coefficient to obtain the detail coefficient cD′ after threshold shrinkage processing. l .

[0069] In the present invention, the specific formula of the soft threshold function is: cD′ l =sgn(cD l )·max(0,|cD l |-δ l ), where cD′ l is the detail coefficient of the lth layer after threshold shrinkage processing, δ l is the threshold corresponding to the detail coefficient of the lth layer, cD l is the detail coefficient corresponding to the lth layer, sgn() is the sign function, max() is the maximum value function, and 0 is to ensure that the minimum value of the coefficient is not negative.

[0070] Step S25: The detail coefficients cD′ of each layer after threshold shrinkage l and the approximation coefficient cA L Perform wavelet reconstruction to obtain the preprocessed signal x c (t), where c∈{X,Y,Z}. The threshold value in the soft threshold function of the present invention can be adaptively calculated according to the Stein unbiased risk estimation criterion to achieve the best denoising effect. In addition, the x obtained after reconstruction c (t) The signal-to-noise ratio is higher and the key fault impulse profile is retained, which will serve as the input of the subsequent feature extraction module.

[0071] Step S3: Perform parallel dual-path analysis on the pre-processed signals corresponding to each axis to obtain the frequency domain and time-frequency domain features corresponding to each axis. The dual paths are divided into a first path and a second path. The first path primarily identifies persistent or periodic vibration components in the signal that may be caused by certain types of joint misalignment faults, such as periodic unevenness, regular impact, or continuous friction, or imbalance of the equipment's own rotating components. The second path primarily detects and characterizes transient impacts or non-stationary changes in the signal caused by impact events in surface protrusions or joint misalignment. These features may be ignored or averaged in traditional FFT (Fast Fourier Transform) analysis.

[0072] Step S3: Performing parallel dual-path analysis on the pre-processed signals corresponding to each axis to obtain the frequency domain features and time-frequency domain features corresponding to each axis, specifically including:

[0073] Frequency domain analysis of the first path: The Welch method is used to calculate the power spectral density of the preprocessed signal, and the power spectral density diagram of each axis is drawn based on the power spectral density. The peak detection algorithm is used on the power spectral density diagram to extract n frequency peaks with maximum energy and the amplitudes corresponding to the frequency peaks as frequency domain features, namely FFT features.

[0074] The first path of the present invention is used to identify steady or quasi-steady vibration signals caused by regular misalignment or periodic friction in the tank joints. The Welch's method is used to calculate the power spectral density (PSD) of the signal to obtain a smoother, more statistically robust and more reliable spectrum estimate than the standard FFT. The specific parameters of the Welch method in this embodiment are set as follows: the window function uses a Hamming window (HammingWindow) that can effectively suppress spectrum leakage, the window length is 1024 sampling points, and the overlap rate (Overlap) between windows is set to 50%. On the calculated power spectral density graph, the peak detection algorithm is applied, and the minimum peak height (MinPeakHeight) is set to 1.5 times the spectrum mean, and the minimum peak distance (MinPeakDistance) is set to 5Hz to avoid detecting irrelevant clutter peaks. Sort by peak energy from high to low, and extract the n most significant frequency peaks, taking n=3. Finally, n groups of frequency peak-amplitude pairs are obtained:

[0075] The present invention applies an advanced peak detection algorithm to the calculated power spectrum density diagram of each axis, based on a preset amplitude threshold, peak prominence, or by accurately interpolating the peak position and amplitude through quadratic polynomial fitting, to identify and extract several groups of frequency peak-amplitude pairs for each axis, including: several most significant frequency peaks (f FXi ,f FYi ,f Fzi ); the amplitude of the corresponding frequency peak (A FXi ,A FYi ,A FZi ).

[0076] Time-frequency domain analysis of the second path: Complex Morley wavelet is used as the mother wavelet, and continuous wavelet transform (CWT) is performed on the preprocessed signal to obtain the time-frequency energy. The transformation formula is:

[0077]

[0078] Where x(t) is the angular velocity signal to be analyzed, that is, the preprocessed signal x c(t), a is the scale factor which is inversely proportional to the frequency and controls the expansion and contraction of the wavelet, b is the translation factor used for time positioning and controls the translation of the wavelet, ψ * is the complex conjugate of the mother wavelet function, W(a,b) is the time-frequency energy;

[0079] The complex Morley wavelet function can be defined as:

[0080]

[0081] Where, f c is the center frequency of the mother wavelet, f b is the bandwidth parameter, Ψ M (t) is the complex Morley wavelet function, f c and f b Together they determine the focusing characteristics of the wavelet in the time-frequency plane.

[0082] Based on the time-frequency energy, a time-frequency energy graph is drawn, and the local energy maximum is searched on the time-frequency energy graph to identify m transient impact events. The occurrence time, center frequency and peak energy corresponding to each transient impact event are used as the time-frequency domain features, namely CWT features.

[0083] The present invention identifies regions with significant energy concentration and localization in time from the time-frequency energy diagram, which generally correspond to transient vibrations caused by faults.

[0084] The second path of the present invention is used to accurately capture and quantify extremely short-duration transient impact events caused by protruding stains on the tank surface, severe impacts on welds or joints. The Complex Morlet wavelet, which has excellent time-frequency aggregation for analyzing impact signals, is selected as the mother wavelet, and the signal is subjected to continuous wavelet transform to generate a time-frequency energy graph that can simultaneously display time and frequency information. On the generated time-frequency graph, transient events with concentrated energy are identified by using blob detection in two-dimensional image processing or by directly searching for local energy maxima. Extract the m events with the strongest energy, and take m=2. For each identified event, accurately extract its peak energy, the center frequency corresponding to the peak energy, and the time when the impact occurs, and finally obtain m groups of energy-frequency-time triplets: The moment when the transient shock event occurs (t CXi ,t CYi ,t CZi ); the center frequency corresponding to the peak energy (f CXi ,f CYi ,f CZi ); Peak energy of transient impact event (E CXi ,E CYi ,E CZi ).

[0085] In the present invention, the first path uses a frequency domain analysis method to extract frequency domain features that can quantify the periodic components of the signal, and the second path uses a time-frequency domain analysis method to extract time-frequency domain features that can quantify the transient impact components of the signal.

[0086] Step S4: The frequency domain features and time-frequency domain features corresponding to the three axes are spliced ​​in a predetermined order to obtain a comprehensive fault feature vector. The specific formula is:

[0087]

[0088] Among them, f F,c,i ,A F,c,i The frequency peak value and amplitude corresponding to the peak value of the i-th peak energy on the c-th axis are sorted from high to low according to the peak energy, E C,c,j ,f C,c,j ,t C,c,j are the peak energy, center frequency and occurrence time corresponding to the jth transient impact time of the cth axis respectively. The dimension of the comprehensive fault feature vector of the present invention is fixed, that is, 3*(m+n).

[0089] The present invention concatenates all the features extracted from the X, Y, and Z axis signals, i.e., the 3×n frequency-amplitude pairs of the first path and the 3×m energy-frequency-time triples of the second path, in the predetermined order of X-axis FFT features, X-axis CWT features, Y-axis FFT features, Y-axis CWT features, Z-axis FFT features, and Z-axis CWT features. A fixed-length comprehensive feature vector V is constructed, i.e.

[0090]

[0091] According to the parameters set in this embodiment (n = 3, m = 2), each axis generates 2 × 3 (FFT) + 3 × 2 (CWT) = 12 features. The three axes generate a total of 12 × 3 = 36 features. The final output feature vector V is a 36-dimensional numerical vector.

[0092] The comprehensive feature vector disclosed in the present invention not only includes the above-mentioned frequency, amplitude, time, energy and other characteristics, but can also be supplemented by introducing other statistical features that can enhance fault differentiation, such as the energy distribution, spectral kurtosis, and skewness of each axis signal in a specific frequency band.

[0093] Fault type identification:

[0094] The characteristic hypothesis of joint misalignment is that, depending on its physical cause, such as periodic unevenness, regular impact, or continuous friction, it may appear in the FFT spectrum as a stable peak at a specific low or medium frequency related to the lifting vessel's speed and the joint spacing; or in the CWT time-frequency diagram as a repetitive transient energy cluster with a certain temporal regularity, such as at equal intervals. The vibration amplitude is generally related to the severity and type of misalignment, such as lateral or vertical misalignment.

[0095] Characteristic Hypothesis for Surface Defect Protrusions: Because these are typically localized, sudden obstacles on the tankway surface, they produce a sharp, short impact when a container is lifted through. Therefore, their primary characteristic is the appearance of a relatively isolated, high-energy transient impact in the CWT time-frequency diagram, with a wide frequency range and, in particular, significant high-frequency components. The FFT spectrum may only show an overall increase in high-frequency energy, but the characteristics of individual impacts may not be prominent. The impact amplitude is related to factors such as the size and shape of the protrusion and the relative velocity of the container during passage.

[0096] Assessment of the main fault direction: By comprehensively analyzing the relative amplitude of the fault characteristics on the X, Y, and Z three-axis gyroscope signals, the order of occurrence, and the combination pattern of the impact energy of each axis, the main rotation direction affected by the fault or the main action axis of the impact torque can be inferred.

[0097] If a strong transient impact energy is mainly concentrated in the X-axis angular velocity signal, while the Y-axis and Z-axis responses are small or delayed, it may indicate a shock event that mainly causes the lifting container to rotate about its X-axis.

[0098] If a persistent, significant characteristic frequency vibration is detected in the FFT spectrum of the Y axis, while there is no such phenomenon or the response is weak in the X and Z axes, it may indicate that there is a joint misalignment that causes periodic disturbances around the Y axis.

[0099] Step S5: Diagnose the fault type of the rigid tankway based on the comprehensive fault feature vector, determine the main direction of abnormal angular movement caused by the fault, and output the diagnosis result. The diagnosis result includes the fault type and the main direction of abnormal angular movement caused by the fault; the fault type is "joint misalignment" or "surface protrusion." There are two ways to diagnose the fault type of the rigid tankway based on the comprehensive fault feature vector. One is to use the constructed comprehensive feature vector V and preset classification logic to diagnose the fault type of the rigid tankway; the other is to use the constructed comprehensive feature vector V and a trained machine learning model to diagnose the fault type of the rigid tankway.

[0100] Among them, the preset classification logic includes: The fault type is determined to be surface protrusion, where T E1 is the high energy threshold, TA1 is the low amplitude threshold, E C,c,j is the peak energy corresponding to the jth transient impact time of the cth axis, that is, the maximum energy value, A F,c,i The amplitude corresponding to the frequency peak corresponding to the peak energy of the i-th peak on the c-th axis is sorted from high to low by peak energy, that is, the maximum amplitude; if The fault type is determined to be joint dislocation, where T A2 is the high amplitude threshold, T E2 is the low energy threshold;

[0101] When a fault is determined, the main direction of abnormal angular motion caused by the fault is determined by the axial direction of the maximum response, and the determination formula is:

[0102] Direction = argmax{KeyFeature c}

[0103] In the formula, KeyFeaturec is the key feature value that causes the judgment, such as E C,c,j or A F,c,i , Direction is the main direction of abnormal angular motion caused by the fault, and argmax{} is the parameter that takes the maximum value of the inner expression.

[0104] The following is an explanation of the above formula:

[0105] Surface stain protrusion diagnosis: If the maximum energy value E extracted by the CWT path C,c,j Exceeding the high energy threshold T E1 (Indicates the presence of a severe impact), and the maximum amplitude A extracted by the FFT path F,c,i Below the low amplitude threshold T A1 (indicating a lack of periodicity), it is determined to be a surface protrusion fault.

[0106] Joint misalignment diagnosis: If the maximum amplitude A extracted by the FFT path F,c,i Exceeding the high amplitude threshold T A2 (Indicates the presence of significant periodic vibration), and the maximum energy value E extracted by the CWT path C,c,j Below the low energy threshold T E2 (Indicating that there is no severe single impact), it is determined to be a joint misalignment fault.

[0107] The fault direction is determined by the axis that causes the maximum response. Example: If it is determined to be a surface protrusion and the energy E on the X axis is C,X If the value is the largest among the three axes, the fault direction is determined to be the X-axis. The system finally outputs a structured diagnostic report.

[0108] The present invention uses a trained machine learning model to diagnose fault types: in order to achieve a higher level of diagnostic accuracy and robustness, especially when dealing with complex working conditions and subtle fault feature differences, a machine learning classifier can be used to train and obtain a machine learning model. The prerequisite is that a sufficient number of sample data with accurate fault type and direction labels can be obtained. These data can be generated by building an experimental platform that simulates tank faults or using a high-fidelity dynamic simulation model. When implementing the machine learning classifier, the extracted original feature vectors need to be preprocessed as necessary, such as feature selection (removing redundant or irrelevant features), feature scaling (such as standardization or normalization), and rigorous model training, cross-validation, and independent test set evaluation to ensure the generalization ability of the model and its reliability in practical applications. The specific process is consistent with the conventional training method and will not be discussed in detail here.

[0109] Through the above steps, the present invention can effectively use gyroscope sensor data to identify the type and direction of faults such as tank channel joint misalignment and tank channel surface stain protrusions in the lifting system tank channel, providing technical support for the safe operation and maintenance decision-making of the lifting system.

[0110] The beneficial effects of the present invention are:

[0111] 1. Improve fault differentiation capabilities: By performing FFT and CWT analysis on the gyroscope's X, Y, and Z axis signals, more comprehensive fault characteristics can be extracted, effectively distinguishing periodic, quasi-periodic, and transient impact characteristics, thereby improving the accuracy of identifying two fault types: joint misalignment and surface protrusion.

[0112] 2. Fault direction assessment: By comprehensively analyzing the response characteristics of the three-axis gyroscope signal, the main direction of the fault impact or the direction of abnormal changes in the cage attitude can be assessed.

[0113] 3. Enhanced diagnostic information dimension: Compared with diagnostic technologies that rely solely on attitude angles or single signal processing methods, the present invention combines frequency domain and time-frequency domain analysis and focuses on multi-axis information, providing richer diagnostic information for tankway health status assessment.

[0114] 4. Optimize feature extraction: Based on the different signal characteristics (stationary periodicity and short-term transients) that may be exhibited by different types of faults, the use of two complementary analysis methods, FFT and CWT, helps to capture fault information more comprehensively.

[0115] The present invention also provides a gyroscope-based vertical shaft rigid tankway fault diagnosis system, the system comprising:

[0116] The acquisition module is used to obtain the original angular velocity signals of the X, Y, and Z axes.

[0117] The preprocessing module is used to preprocess the original angular velocity signal of each axis respectively to obtain a preprocessed signal corresponding to each axis.

[0118] The dual-path analysis module is used to perform parallel dual-path analysis on the pre-processed signals corresponding to each axis to obtain the frequency domain characteristics and time-frequency domain characteristics corresponding to each axis.

[0119] The splicing module is used to splice the frequency domain features and time-frequency domain features corresponding to the three axes in a predetermined order to obtain a comprehensive fault feature vector.

[0120] The fault diagnosis module is used to diagnose the fault type of the rigid tank channel according to the comprehensive fault characteristic vector, determine the main direction of abnormal angular movement caused by the fault, and output the diagnosis result.

[0121] As an optional implementation manner, the dual-path analysis module of the present invention includes:

[0122] The frequency domain analysis unit of the first path is used to calculate the power spectral density of the preprocessed signal using the Welch method, and draw a power spectral density diagram of each axis based on the power spectral density, and extract n frequency peaks with maximum energy and the amplitudes corresponding to the frequency peaks as frequency domain features, i.e., FFT features, on the power spectral density diagram through a peak detection algorithm; wherein n is a positive integer greater than 1.

[0123] The time-frequency domain analysis unit of the second path is used to use the complex Morlay wavelet as the mother wavelet, perform continuous wavelet transform (CWT) processing on the preprocessed signal to obtain time-frequency energy, draw a time-frequency energy graph based on the time-frequency energy, search for local energy maxima on the time-frequency energy graph to identify m transient impact events, and use the occurrence time, center frequency and peak energy corresponding to each transient impact event as time-frequency domain features, i.e., CWT features; wherein m is a positive integer greater than 1.

[0124] For the parts that are the same as the method, please refer to the method and no further discussion will be given here.

[0125] The solution disclosed in the present invention is particularly suitable for detecting common fault types such as tank channel joint dislocation and tank channel surface defects in the tank channel of the lifting system, and determining the main direction of abnormal angular movement caused by the fault.

[0126] Based on the above Figures 1 to 2 The method shown in FIG. 1 is a method for performing the above-mentioned steps. Accordingly, this embodiment further provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the computer program can realize the above-mentioned steps. Figures 1 to 2 The method shown.

[0127] Based on the above Figures 1 to 2The method shown in FIG. 1 is a method for performing the above-mentioned operations. Accordingly, this embodiment further provides a computer program product having a computer program stored thereon. When the computer program is executed by a processor, the above-mentioned Figures 1 to 2 The method shown.

[0128] Based on this understanding, the technical solution of the present application can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, USB flash drive, mobile hard disk, etc.), and includes a number of instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute the methods of various implementation scenarios of the present application.

[0129] Based on the above Figures 1 to 2 In order to achieve the above purpose, the present invention also provides an electronic device that can be configured on a computer terminal, etc. The device includes a storage medium and a processor; the storage medium is used to store a computer program; the processor is used to execute the computer program to achieve the above purpose. Figures 1 to 2 The method shown.

[0130] Based on the above Figures 1 to 2 In order to achieve the above purpose, the embodiment of the present application further provides a chip, including one or more interface circuits and one or more processors; the interface circuit is used to receive a signal from the memory of the electronic device and send the signal to the processor, the signal including the computer instruction stored in the memory; when the processor executes the computer instruction, the electronic device executes the above Figures 1 to 2 The method shown.

[0131] Optionally, the physical device may further include a user interface, a network interface, a camera, a radio frequency (RF) circuit, a sensor, an audio circuit, a Wi-Fi module, and the like. The user interface may include a display, an input unit such as a keyboard, and the like. The optional user interface may also include a USB interface, a card reader interface, and the like. The network interface may optionally include a standard wired interface, a wireless interface (such as a Wi-Fi interface), and the like.

[0132] Those skilled in the art will understand that the above-mentioned physical device structure provided in this embodiment does not constitute a limitation on the physical device, and may include more or fewer components, or a combination of certain components, or different component arrangements.

[0133] The storage medium may also include an operating system and a network communication module. The operating system is a program that manages the hardware and software resources of the physical device, supporting the execution of information processing programs and other software and / or programs. The network communication module is used to enable communication between components within the storage medium, as well as with other hardware and software within the physical information processing device.

[0134] Through the description of the above implementation methods, those skilled in the art can clearly understand that the present application can be implemented by means of software plus a necessary general hardware platform, or by hardware.

[0135] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprises" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device that includes a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a..." do not exclude the presence of other identical elements in the process, method, article or device that includes the elements.

[0136] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. Reference can be made to the common and similar parts between the various embodiments. For the systems disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple, and the relevant parts can be referred to the method description.

[0137] This document uses specific examples to illustrate the principles and implementation methods of the present invention. The above examples are only intended to help understand the method and core concept of the present invention. At the same time, those skilled in the art will find that the specific implementation methods and application scopes may vary based on the concept of the present invention. In summary, the contents of this specification should not be construed as limiting the present invention.

Claims

1. A gyroscope-based fault diagnosis method for a vertical shaft rigid tankway, characterized in that: The method comprises: Get the original angular velocity signals of X, Y and Z axes; Preprocessing the original angular velocity signal of each axis respectively to obtain a preprocessed signal corresponding to each axis; Performing parallel dual-path analysis on the pre-processed signals corresponding to each axis to obtain frequency domain features and time-frequency domain features corresponding to each axis; The frequency domain features and time-frequency domain features corresponding to the three axes are spliced ​​in a predetermined order to obtain a comprehensive fault feature vector; The fault type of the rigid tank channel is diagnosed according to the comprehensive fault characteristic vector, and the main direction of the abnormal angular motion caused by the fault is determined, and the diagnosis result is output.

2. The gyroscope-based vertical shaft rigid tankway fault diagnosis method according to claim 1 is characterized in that: Preprocessing the original angular velocity signal of each axis to obtain a preprocessed signal corresponding to each axis specifically includes: Daubechies 4 wavelet is selected as the mother wavelet, and the original angular velocity signal x of each axis is raw,c (t) Perform L-layer wavelet decomposition to obtain the detail coefficients cD of each layer l and the approximation coefficient cA L ; Calculate the noise standard deviation based on the median estimation method; Determine the threshold corresponding to the detail coefficient of each layer according to the noise standard deviation; A soft threshold function with a threshold corresponding to each layer detail coefficient is used to perform threshold shrinkage processing on each layer detail coefficient to obtain the detail coefficient cD after threshold shrinkage processing. l '; The detail coefficients cD of each layer after the threshold shrinkage process l ′ and approximation coefficient cA L Perform wavelet reconstruction to obtain the preprocessed signal x c (t).

3. The gyroscope-based vertical shaft rigid tankway fault diagnosis method according to claim 2 is characterized in that: The specific formula of the soft threshold function is: CD' l =sgn(cD l )·max(0,|cD l |-d l ) Among them, cD′ l is the detail coefficient of the lth layer after threshold shrinkage processing, δ l is the threshold corresponding to the detail coefficient of the lth layer, cD l is the detail coefficient corresponding to the lth layer, sgn() is the sign function, max() is the maximum value function, and 0 is to ensure that the minimum value of the coefficient is not negative.

4. The gyroscope-based vertical shaft rigid tankway fault diagnosis method according to claim 1 is characterized in that: The performing of parallel dual-path analysis on the pre-processed signals corresponding to the respective axes to obtain frequency domain features and time-frequency domain features corresponding to the respective axes specifically includes: Frequency domain analysis of the first path: The Welch method is used to calculate the power spectral density of the preprocessed signal, and a power spectral density graph of each axis is plotted based on the power spectral density. A peak detection algorithm is then used to extract n frequency peaks with maximum energy and their corresponding amplitudes from the power spectral density graph as frequency domain features, i.e., FFT features. Here, n is a positive integer greater than 1. Time-frequency domain analysis of the second path: complex Morlay wavelet is used as the mother wavelet, and continuous wavelet transform (CWT) processing is performed on the preprocessed signal to obtain time-frequency energy. A time-frequency energy graph is drawn based on the time-frequency energy, and local energy maxima are searched on the time-frequency energy graph to identify m transient impact events. The occurrence time, center frequency and peak energy corresponding to each transient impact event are used as time-frequency domain features, namely CWT features; where m is a positive integer greater than 1.

5. The gyroscope-based vertical shaft rigid tankway fault diagnosis method according to claim 1 is characterized in that: The frequency domain features and time-frequency domain features corresponding to the three axes are spliced ​​in a predetermined order to obtain a comprehensive fault feature vector. The specific formula is: Among them, f F,c,i ,A F,c,i The frequency peak value and amplitude corresponding to the peak value of the i-th peak energy on the c-th axis are sorted from high to low according to the peak energy, respectively. C,c,j ,f C,c,j ,t C,c,j are the peak energy, center frequency and occurrence time corresponding to the jth transient impact time on the cth axis, respectively. c∈{X,Y,Z}, where X,Y,Z are the three axes.

6. The gyroscope-based vertical shaft rigid tankway fault diagnosis method according to claim 5 is characterized in that: The diagnosing the fault type of the rigid tank channel according to the comprehensive fault feature vector specifically includes: The fault type of the rigid tank channel is diagnosed by using the constructed comprehensive feature vector and a preset classification logic; Alternatively, the fault type of the rigid tank channel may be diagnosed by using the constructed comprehensive feature vector and a trained machine learning model.

7. A gyroscope-based vertical shaft rigid tankway fault diagnosis system, characterized in that: The system comprises: The acquisition module is used to obtain the original angular velocity signals of the X, Y, and Z axes; A preprocessing module, configured to preprocess the original angular velocity signal of each axis to obtain a preprocessed signal corresponding to each axis; A dual-path analysis module is used to perform parallel dual-path analysis on the pre-processed signals corresponding to each axis to obtain frequency domain features and time-frequency domain features corresponding to each axis; A splicing module is used to splice the frequency domain features and time-frequency domain features corresponding to the three axes in a predetermined order to obtain a comprehensive fault feature vector; The fault diagnosis module is used to diagnose the fault type of the rigid tank channel according to the comprehensive fault characteristic vector, determine the main direction of abnormal angular movement caused by the fault, and output the diagnosis result.

8. The gyroscope-based vertical shaft rigid tankway fault diagnosis system according to claim 7 is characterized in that: The dual-path analysis module includes: The frequency domain analysis unit of the first path is used to calculate the power spectral density of the preprocessed signal using the Welch method, draw a power spectral density graph of each axis based on the power spectral density, and extract n frequency peaks with maximum energy and the amplitudes corresponding to the frequency peaks as frequency domain features, i.e., FFT features, from the power spectral density graph using a peak detection algorithm; wherein n is a positive integer greater than 1; The time-frequency domain analysis unit of the second path is used to use the complex Morlay wavelet as the mother wavelet, perform continuous wavelet transform (CWT) processing on the preprocessed signal to obtain time-frequency energy, draw a time-frequency energy graph based on the time-frequency energy, search for local energy maxima on the time-frequency energy graph to identify m transient impact events, and use the occurrence time, center frequency and peak energy corresponding to each transient impact event as time-frequency domain features, i.e., CWT features; wherein m is a positive integer greater than 1.

9. An electronic device comprising a storage medium, a processor, and a computer program stored on the storage medium and executable on the processor, wherein: When the processor executes the computer program, the method according to any one of claims 1 to 6 is implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 6 is implemented.

Images