Switch machine fault diagnosis system and method based on spectral analysis

Abstract

The invention discloses a point switch fault diagnosis system and method based on spectral analysis, and relates to the technical field of point switch fault diagnosis, and the system comprises the following components: a data collection module which selects high-precision acceleration, vibration, load, temperature and humidity sensors, collects the frequency spectrum data and environmental parameters of the operation of a point switch in real time, and transmits the frequency spectrum data and environmental parameters to a server; data are transmitted in a wired or wireless communication mode; by adopting the multi-scale spectrum analysis method, the spectrum data of the switch machine can be comprehensively analyzed from different time and frequency scales, the spectrum change trend can be analyzed on a long-time scale, the aging and abrasion conditions of equipment can be judged, high-frequency detail analysis can be carried out on a short-time scale, sudden fault events can be found in time, and the safety of the switch machine is improved. Meanwhile, signals in different frequency ranges are analyzed on the frequency scale, fault information of different levels is obtained, the fault information in the operation process of the switch machine is comprehensively obtained, and the accuracy of fault diagnosis is remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field of switch machine fault diagnosis, and specifically to a switch machine fault diagnosis system and method based on spectrum analysis. Background Technique

[0002] In the railway signal system, as a key device, the operating state of the switch machine is closely related to the safety of railway operation. During the actual operation process, the switch machine will inevitably have various faults due to various factors such as mechanical wear and electrical failures. Therefore, it is crucial to conduct effective fault diagnosis on the switch machine. Only by timely and accurately detecting faults can the safe and stable operation of the railway system be ensured.

[0003] Traditional switch machine fault diagnosis methods mainly rely on the monitoring of some operating parameters. However, this method has obvious limitations. On the one hand, traditional spectrum analysis methods usually analyze signals at a single scale and cannot simultaneously take into account the macroscopic trend and microscopic abnormal changes of the switch machine operation. This leads to the easy omission of important fault information during the fault diagnosis process, thereby affecting the accuracy and timeliness of fault diagnosis. On the other hand, the parameter settings of traditional spectrum analysis are fixed and lack flexibility. Since the switch machine operates under different working conditions such as load, temperature, and humidity, its fault characteristic frequencies will also change. The traditional spectrum analysis method with fixed parameters cannot adapt to these changes and is difficult to accurately capture the fault characteristic frequencies, resulting in insufficient sensitivity of fault diagnosis and being unable to effectively meet the actual needs of switch machine fault diagnosis.

[0004] In view of the above problems, it is necessary to optimize the existing switch machine fault diagnosis system and method. By adopting an adaptive spectrum analysis algorithm and a multi-scale spectrum analysis method, it is possible to achieve comprehensive monitoring of the switch machine operating state and accurate diagnosis of faults. Therefore, it is of great significance to develop a switch machine fault diagnosis system and method based on spectrum analysis that can comprehensively achieve the above characteristics. Summary of the Invention

[0005] The object of the present invention is to make up for the deficiencies of the prior art and provide a switch machine fault diagnosis system and method based on spectrum analysis. It can comprehensively monitor the operating state of the switch machine and accurately diagnose faults by adopting an adaptive spectrum analysis algorithm and a multi-scale spectrum analysis method. Among them, the adaptive spectrum analysis algorithm can automatically adjust parameters according to the real-time operating state and environmental changes of the switch machine, enabling the spectrum analysis to more accurately capture the fault characteristic frequencies, enhancing the adaptability to different working conditions. The multi-scale spectrum analysis method can analyze the spectrum data of the switch machine from different time and frequency scales, comprehensively obtain the fault information during the operation of the switch machine, and improve the accuracy of fault diagnosis. In addition, the fault diagnosis and feature extraction module of the present invention simplifies the system structure and improves the fault diagnosis efficiency. The system can timely detect sudden faults and potential hidden dangers of the switch machine, provide strong support for equipment preventive maintenance, and has strong versatility and scalability, and can be applied to the fault diagnosis of different types of switch machines and other similar equipment.

[0006] To solve the above technical problems, the present invention provides the following technical solutions: On the one hand, a switch machine fault diagnosis system based on spectrum analysis, the system includes the following components:

[0007] Data acquisition module: High-precision acceleration, vibration, load, temperature, and humidity sensors are selected to collect the spectrum data and environmental parameters of the switch machine in real time, and transmit the data through wired or wireless communication methods;

[0008] Adaptive spectrum analysis module: According to the load, temperature, and humidity parameters transmitted by the data acquisition module, an algorithm for dynamically adjusting the frequency band range and resolution is used to analyze and process the collected spectrum data to capture the fault characteristic frequencies;

[0009] Multi-scale spectrum analysis module: The time scale is divided into long time scale and short time scale, the frequency is divided into different ranges, the spectrum characteristics are analyzed respectively and the results are fused. The equipment aging and wear are judged through the long time scale, the sudden faults are captured through the short time scale, and the fault information at different levels is analyzed through the frequency scale, so as to comprehensively master the operating state and fault conditions of the switch machine;

[0010] Fault diagnosis and feature extraction module: Extract the peak value, mean value, variance, and frequency offset fault characteristic parameters of the spectrum from the multi-scale spectrum analysis results and fuse them, compare with the preset threshold to judge the fault, determine the fault type and location, record and analyze the fault information, and provide data support for maintenance prevention;

[0011] Data storage module: Adopt a database management system, allocate storage capacity according to the data volume and storage ratio, classify and store various types of data, establish a backup and recovery mechanism, and regularly perform data management and maintenance to ensure the operation of the system and the availability of data.

[0012] Furthermore, the adaptive spectrum analysis module receives the turnout load, temperature, and humidity environmental parameters transmitted by the data acquisition module, monitors and analyzes these parameters in real time, establishes an association model between the environmental parameters and the spectrum analysis parameters, and dynamically adjusts the resolution and frequency band range parameters of the spectrum analysis according to the analysis results of the environmental parameters. When the load increases, the resolution of the spectrum analysis is adjusted to observe the spectrum details and capture the fault characteristic frequencies that may be generated due to the load change. If the environmental temperature or humidity exceeds the normal range, the frequency band range is adjusted accordingly to cover the frequency range affected by the environment. The adjusted parameters are used to analyze and process the collected spectrum data. Using the spectrum analysis method, the vibration signal in the time domain is converted into a frequency domain signal to obtain spectrum information. During the analysis process, the spectrum is preprocessed according to different application scenarios and requirements.

[0013] Furthermore, the adaptive spectrum analysis module dynamically adjusts the resolution and frequency band range parameters of the spectrum analysis according to the analysis results of the environmental parameters. The resolution adjustment formula is: Where R new is the adjusted spectrum analysis resolution, R old is the spectrum analysis resolution before adjustment, β is the resolution adjustment coefficient, a constant determined by experiments, used to control the degree of resolution adjustment, n is the number of energy data points collected within a certain time window, and ΔE i is the difference between the turnout operation energy and the average energy at the i-th time point, and the calculation formula is: Where E i is the energy value at the i-th time point, is the average energy value within this time window, and E ref is the reference energy value, a standard value preset according to the energy consumption during the normal operation of the turnout. The frequency band range parameter adjustment formula is: Where, fmin new is the adjusted lower frequency limit of the frequency band, fmax new is the adjusted upper frequency limit of the frequency band, fmin old is the lower frequency limit of the frequency band before adjustment, fmax old is the upper frequency limit of the frequency band before adjustment. α1, α2, and α3 are the coefficients of the influence of load, temperature, and humidity on the frequency band range, obtained through a large number of experiments and data analyses, used to quantify the effect of environmental factors on the frequency band adjustment. ΔL is the difference between the current load and the standard load, ΔT is the difference between the current temperature and the standard temperature, and ΔH is the difference between the current humidity and the standard humidity.

[0014] Furthermore, the ΔL is the difference between the current load and the standard load, and its calculation formula is: ΔL = L - L std, where L is the current load value, and L std is the standard load value, and ΔT is the difference between the current temperature and the standard temperature. Its calculation formula is: ΔT = T - T std , where T is the current temperature value, and T std is the standard temperature value, and ΔH is the difference between the current humidity and the standard humidity. Its calculation formula is: ΔH = H - H std , where H is the current humidity value, and H std is the standard humidity value.

[0015] Furthermore, the multi-scale spectrum analysis module divides the time scale into long time scale and short time scale. The long time scale is used to analyze the macroscopic trend of the switch machine operation. Under the long time scale, statistical analysis is performed on the spectrum data, and the mean and variance statistics of the spectrum are calculated. By observing the change of the statistics over time, the aging and wear trend of the equipment is judged. The short time scale is used to capture the instantaneous abnormal changes during the operation of the switch machine. Under the short time scale, high-frequency detail analysis is performed on the spectrum data, and signal processing methods are used to analyze the spectrum characteristics of each segment of the signal to discover sudden fault events. On the frequency scale, the spectrum is divided into different frequency ranges, and the signals in each frequency range are analyzed separately. According to the structure and working principle of the switch machine, the potential relationship between different frequency ranges and equipment faults is determined. Through the analysis of the signals in different frequency ranges, the spectrum characteristics that can reflect the fault information of different levels of the equipment are extracted. The analysis results of the time scale and the frequency scale are fused, and the spectrum characteristics under different time and frequency scales are comprehensively considered to more comprehensively understand the operation state and fault conditions of the switch machine. At the same time, the spectrum characteristics of different frequency ranges are combined with the analysis results of the time scale to determine the specific location and type of the fault.

[0016] Furthermore, the multi-scale spectrum analysis module divides the time scale into long time scale and short time scale. The long time scale is used to analyze the macroscopic trend of the switch machine operation. Its analysis formula is: where, T trend is the quantization value of the spectrum change trend under the long time scale, which is used to evaluate the long-term state change of equipment aging and wear, S i is the spectrum characteristic value at the i-th time point, S i+1 is the spectrum characteristic value at the (i + 1)-th time point, m is the total number of time points within the long time scale, and Δt i is the i-th time interval, that is, the time difference between two adjacent time points.

[0017] Furthermore, the multi-scale spectrum analysis module divides the time scale into long time scale and short time scale. The short time scale is used to capture the instantaneous abnormal changes during the operation of the switch machine. Its formula is: where, D detailIt is the quantification value of the prominence of high-frequency details on a short time scale, used to capture instantaneous abnormal changes. It is the eigenvalue of the j-th high-frequency component within a short time scale. It is the eigenvalue of the k-th high-frequency component within a short time scale, and p is the number of high-frequency components within a short time scale. w j It is the weight coefficient of the j-th high-frequency component, set according to the importance of the high-frequency component in fault diagnosis. The higher the importance, the greater the weight.

[0018] Furthermore, the fault diagnosis and feature extraction module extracts the peak value, mean value, variance, and frequency offset fault feature parameters of the spectrum from the multi-scale spectrum analysis results and fuses them. The formula is: F composite =γ1×F peak +γ2×F mean +γ3×F var +γ4×F offset , where F composite is the comprehensive fault feature parameter, used to more comprehensively describe the fault state of the switch machine. F peak is the peak value of the spectrum, reflecting the amplitude size of the frequency position where the energy is concentrated in the spectrum. F mean is the mean value of the spectrum, reflecting the average energy level of the spectrum. F var is the variance of the spectrum, measuring the degree of dispersion of the spectrum energy distribution. F offset is the frequency offset of the spectrum, indicating the deviation between the actual spectrum frequency and the standard frequency. γ1, γ2, γ3, γ4 are the weight coefficients of the peak value, mean value, variance, and frequency offset in the comprehensive fault feature parameter, respectively, obtained through a large number of fault sample analyses and machine learning algorithm optimizations, and used to determine the contribution degree of each feature parameter to fault diagnosis.

[0019] On the other hand, a switch machine fault diagnosis method based on spectrum analysis, the method includes the following specific steps:

[0020] Data acquisition: Use the data acquisition module to collect the spectrum data, as well as load, temperature, and humidity parameters during the operation of the switch machine, and perform data transmission.

[0021] Adaptive spectrum analysis: Through the adaptive spectrum analysis module, according to the collected environmental parameters, use the adaptive spectrum analysis algorithm to dynamically adjust the resolution and frequency band range parameters of the spectrum analysis, and perform adaptive spectrum analysis processing on the collected spectrum data.

[0022] Multi-scale spectrum analysis: Use the multi-scale spectrum analysis module to analyze the spectrum data after adaptive spectrum analysis, extract spectrum features at different time scales and frequency scales respectively. Among them, analyze the spectrum change trend at a long time scale and calculate relevant statistics to judge the aging and wear condition of the equipment, and conduct high-frequency detail analysis at a short time scale;

[0023] Fault diagnosis and feature extraction: According to the multi-scale spectrum analysis results, use the fault diagnosis and feature extraction module to extract fault feature parameters, and perform dimensionality reduction processing. Compare the processed fault feature parameters with the pre-set fault thresholds to judge whether there is a fault in the switch machine. If there is a fault, determine the fault type and location, send an alarm signal, and record the fault information in the data storage module;

[0024] Data storage: Store the collected spectrum data, environmental parameters, adaptive spectrum analysis results, multi-scale spectrum analysis results, fault feature parameters, and fault diagnosis result information in the data storage module for subsequent analysis and query.

[0025] Compared with the prior art, the switch machine fault diagnosis system and method based on spectrum analysis have the following beneficial effects:

[0026] First, by adopting the multi-scale spectrum analysis method, the present invention can comprehensively analyze the spectrum data of the switch machine from different time and frequency scales. Analyze the spectrum change trend at a long time scale to judge the aging and wear condition of the equipment, conduct high-frequency detail analysis at a short time scale to detect sudden fault events in time, and analyze signals in different frequency ranges at the frequency scale to obtain fault information at different levels. Comprehensively obtain the fault information during the operation of the switch machine, significantly improve the accuracy of fault diagnosis, help to more accurately grasp the operation state of the switch machine, and ensure the safety of railway operation.

[0027] Second, through the adaptive spectrum analysis algorithm, the present invention can automatically adjust parameters according to the real-time operation state and environmental changes of the switch machine, and dynamically adjust parameters such as the resolution and frequency band range of spectrum analysis according to environmental factors such as load, temperature, and humidity to achieve accurate capture of fault characteristic frequencies, enhance the adaptability of the system to different working conditions, enable spectrum analysis to more accurately capture fault characteristic frequencies, overcome the disadvantages of fixed parameter settings in traditional spectrum analysis and difficulty in adapting to working condition changes, effectively improve the sensitivity of fault diagnosis, and thus can more timely detect potential faults of the switch machine, providing strong support for preventive maintenance and fault prevention of the equipment.

[0028] Other advantages, objects, and features of the present invention will be set forth in part in the following description, and in part will be obvious to those skilled in the art upon examination of the following, or may be learned from the practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or in the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present invention, and for those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0030] Figure 1 It is an operation flowchart of a switch machine fault diagnosis system and method based on spectrum analysis;

[0031] Figure 2 It is a flowchart of a switch machine fault diagnosis system and method based on spectrum analysis. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0032] To further elaborate on the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following will, in conjunction with the accompanying drawings and preferred embodiments, describe in detail the specific embodiments, structures, features, and their effects of the present invention as follows.

[0033] Embodiment 1

[0034] At multiple busy stations on a certain ordinary railway main line, where the daily train traffic volume at these stations is large and the switch machines are used frequently, in order to ensure the safety and smoothness of railway operation, the switch machine fault diagnosis system based on spectrum analysis of the present invention is installed.

[0035] On each switch machine, high-precision acceleration sensors, vibration sensors, load sensors, temperature sensors, and humidity sensors are carefully installed. The initial sampling frequency is set to once per minute, but it will be dynamically adjusted according to the difference between the actual operating speed of the switch machine and the reference speed. For example, at a certain moment, a heavy-haul train passes by, and the switch machine needs to bear a greater load and act quickly. At this time, its operating speed is higher than the reference speed. The data acquisition module adjusts the sampling frequency dynamically according to the data sampling frequency adjustment strategy and quickly increases the sampling frequency to once every 30 seconds to ensure that the vibration spectrum data of the switch machine under heavy load and high-speed operation, as well as environmental parameters such as load, temperature, and humidity, can be collected more densely. The collected data is transmitted in real time through a stable wireless communication module.

[0036] After receiving the collected data, the adaptive spectrum analysis module monitors and deeply analyzes environmental parameters such as load, temperature, and humidity in real time. When it detects that the environmental temperature rises, for example, on a hot summer afternoon, the temperature is 10°C higher than the normal operating temperature, and at the same time the load increases, such as when multiple heavy-haul trains pass continuously, the formula for adaptively adjusting the frequency band range and as well as the formula for adaptively adjusting the resolution are used for corresponding adjustments. Among them, is the lower limit frequency of the adjusted frequency band, is the upper limit frequency of the adjusted frequency band, is the lower limit frequency of the frequency band before adjustment, is the upper limit frequency of the frequency band before adjustment. α1, α2, and α3 are the coefficients of the influence of load, temperature, and humidity on the frequency band range (initially set as α1 = 0.5, α2 = 0.3, α3 = 0.2). ΔL is the difference between the current load and the standard load (ΔL = L - L std ), ΔT is the difference between the current temperature and the standard temperature (ΔT = T - T std ), ΔH is the difference between the current humidity and the standard humidity (ΔH = H - H std ), R new is the spectrum analysis resolution after adjustment, R old is the spectrum analysis resolution before adjustment, β is the weight of the influence of energy change on resolution adjustment (initially set as 0.2), n is the number of energy data points collected within a certain time window, ΔE i is the difference between the energy of the switch machine operation and the average energy at the i-th time point E i is the energy value at the i-th time point, is the average energy value within this time window, E ref is the reference energy value. Assuming L std = 500N, the current load L = 700N, ΔL = 700 - 500 = 200N, T std = 25°C, the current temperature T = 35°C, ΔT = 35 - 25 = 10°C, H std = 50%, the current humidity H = 55%, ΔH = 55% - 50% = 5%, then (the actual lower limit takes 0Hz), that is, the frequency band range of the spectrum analysis is correspondingly expanded. At the same time, if n = 10 within this time window, R old = 1Hz, E ref = 100J, then The resolution is improved. Through these adjustments, it is possible to more accurately analyze the spectral data and capture the fault characteristic frequencies that may occur under high-temperature and heavy-load conditions.

[0037] On a long time scale, at the end of each day, the system automatically applies the long-time scale spectral trend quantization formula. The time scale is divided into a long time scale in days and a short time scale in seconds. Analyze the changing trend of the spectral data for the day. Among them, T trend is the quantization value of the changing trend of the spectrum under the long time scale, S i is the spectral characteristic value (such as spectral mean, amplitude of a specific frequency component, etc.) at the i-th time point, S i+1 is the spectral characteristic value at the (i + 1)-th time point, m is the total number of time points within the long time scale, and Δt i is the i-th time interval, that is, the time difference between two adjacent time points. For example, assuming one hour as a time point within a day, m = 24, and Δt i = 1 hour (converted to 3600 seconds). By calculating statistical quantities such as the mean and variance of the spectrum and comparing with historical data, judge the aging and wear conditions of the switch machine. After monitoring for a period of time, it is found that the spectral mean of a certain switch machine has gradually increased by 15% within a week, and the variance also shows an obvious upward trend. Combining past experience and data analysis, it is initially judged that some mechanical components of the switch machine may be worn. On a short time scale, perform high-frequency detail analysis on the spectral data before and after each switch operation, and use the short-time scale high-frequency detail highlighting formula. Among them, D detail is the quantization value of the highlighting degree of the high-frequency details under the short time scale. is the characteristic value (such as amplitude, phase, etc.) of the j-th high-frequency component within the short time scale, p is the number of high-frequency components within the short time scale, and w j is the weight coefficient of the j-th high-frequency component (for high-frequency components related to key component failures, initially set w j = 0.3, for other secondary high-frequency components, set w j = 0.1, etc.). When the switch machine performs a normal conversion operation, the system will perform high-frequency sampling and analysis on the spectral data within a few seconds before and after the operation, and promptly detect sudden fault events such as momentary jamming and loosening of mechanical components. At the same time, on the frequency scale, conduct a detailed analysis of signals in different frequency ranges to determine the relationship between low-frequency signals (0 - 100 Hz) and mechanical structure vibrations, and the association between high-frequency signals (above 1000 Hz) and electrical system abnormalities. For example, when an abnormal signal peak is detected in the high-frequency band, further check the electrical system for electromagnetic interference or electrical component failures.

[0038] According to the results of multi-scale spectrum analysis, fault feature parameters such as the peak value, mean value, variance, and frequency offset of the spectrum are extracted, and these feature parameters are dimensionally reduced to reduce the data dimension and improve the efficiency and accuracy of fault diagnosis. The comprehensive extraction formula for fault feature parameters F composite = γ1×F peak + γ2×F mean + γ3×F var + γ4×F offset is adopted, where F composite is the comprehensive fault feature parameter, F peak is the peak value of the spectrum, F mean is the mean value of the spectrum, F var is the variance of the spectrum, F offset is the frequency offset of the spectrum, and γ1, γ2, γ3, and γ4 are the weight coefficients of the peak value, mean value, variance, and frequency offset in the comprehensive fault feature parameter (initially set as γ1 = 0.4, γ2 = 0.2, γ3 = 0.2, γ4 = 0.2). The processed feature parameters are compared with the pre-set fault threshold. If it is found in a certain analysis that the spectrum peak exceeds the threshold by 20%, and the frequency offset exceeds the normal range by 15%, and at the same time the mean value and variance also deviate from the normal interval, it is judged that the switch machine has a fault. Further, through in-depth analysis of the fault feature parameters and combining the correlation model between the fault feature parameters and different fault types, it is determined that the fault type is mechanical component wear, and the specific location may be a certain key transmission component of the switch machine. The system immediately issues an alarm signal to notify the railway maintenance personnel to conduct inspections and repairs. At the same time, information such as the time of the fault occurrence, feature parameters, and environmental parameters is recorded in detail to provide comprehensive data support for subsequent maintenance and fault prevention. After receiving the alarm, the maintenance personnel can quickly locate the fault location by viewing the detailed information recorded by the system, and carry the corresponding repair tools and components to the site for repair, greatly shortening the fault handling time.

[0039] All the collected data, including spectrum data, environmental parameters, adaptive spectrum analysis results, multi-scale spectrum analysis results, fault feature parameters, and fault diagnosis results, etc., are stored in a relational database. To ensure the security and reliability of the data, the data is regularly backed up, and the backup data is stored in multiple different storage media and separately stored in different geographical locations. At the same time, the storage capacity is dynamically adjusted according to the data volume and storage ratio. When it is found that the storage volume of a certain type of data grows rapidly, the corresponding storage resources are added in a timely manner. In addition, the database is regularly optimized and maintained to clean up expired and useless data, improving the query and retrieval efficiency of the data. After a period of operation, the system successfully detected multiple potential faults of the switch machine, such as early wear of mechanical components and slight abnormalities in the electrical system, etc. By performing maintenance in advance, railway operation accidents caused by switch machine failures were avoided, effectively improving the operation safety and reliability of the railway main line. At the same time, through the analysis of historical data, the railway department can also summarize the occurrence rules of switch machine failures, formulate more reasonable maintenance plans, and further reduce the maintenance cost and equipment failure rate.

[0040] Embodiment 2

[0041] In a key section of a high-speed railway, this section is in a complex terrain area, including multiple viaducts and tunnels, and the train running speed is high and the density is large, with extremely high requirements for the reliability and stability of the switch machine. To ensure the safe and efficient operation of the high-speed railway, the fault diagnosis system of the present invention is deployed here.

[0042] To adapt to the severe characteristics of the high-speed operation of the high-speed railway switch machine, acceleration sensors, vibration sensors, etc. with higher precision and faster response speed are specially selected. The initially set sampling frequency is relatively high, 10 times per second, but it will be flexibly adjusted according to the dynamic changes in the actual running speed of the switch machine. For example, when a high-speed train with a speed of 350 km / h is about to pass and the switch machine is performing a rapid conversion operation, the speed changes extremely rapidly and greatly. At this time, the data acquisition module adjusts the sampling frequency dynamically according to the data sampling frequency adjustment strategy, and quickly increases the sampling frequency to 50 times per second within a very short time, so as to accurately collect the vibration spectrum data of the switch machine under high-speed and high-load operating conditions, as well as environmental parameters such as load, temperature, and humidity. The collected data is transmitted through a high-speed and stable wired communication network.

[0043] After receiving the collected data, the adaptive spectrum analysis module immediately monitors and deeply analyzes environmental parameters such as load, temperature, and humidity in real time and with high precision. Since high-speed railways have extremely high requirements for the operating accuracy of switch machines, even if there are slight changes in environmental parameters, this module can quickly respond. For example, when it is detected that the humidity has increased by 15% within a short period of time and the load is at a high level (more than 80% of the rated load), the adaptive spectrum analysis module quickly adjusts the spectrum analysis frequency band range and resolution according to the adaptive adjustment strategies for the frequency band range and resolution. It specifically expands the frequency band range of spectrum analysis, lowers the lower limit of the frequency band by 80 Hz, and raises the upper limit by 150 Hz. At the same time, it further improves the resolution from the original 0.5 Hz precision to 0.2 Hz precision. Through these fine adjustments, it is possible to conduct a more detailed and comprehensive analysis of spectrum data, effectively capture the fault characteristic frequencies that may affect the performance of the switch machine under complex environments and high-load working conditions, and not miss any potential fault risk points.

[0044] The multi-scale spectrum analysis module sets the time scales as a long time scale in months and a short time scale in milliseconds. On the long time scale, at the end of each month, the system automatically conducts a comprehensive and in-depth trend analysis of the spectrum data for that month. By calculating various statistics such as the mean, variance, and frequency distribution of the spectrum, and making a detailed comparison with historical data and the standard data of the same type of switch machine, it accurately evaluates the long-term operating status and aging degree of the switch machine. For example, through continuous monitoring for several months, it is found that the spectrum mean of a certain switch machine has gradually increased by 12% within three months, and the variance also shows a slow upward trend. Combining professional data analysis models and expert experience, it is initially judged that some key components of this switch machine may have early wear or performance degradation. On the short time scale, it conducts high-frequency and high-precision detailed analysis of the spectrum data for high-speed switch operations (usually completed within milliseconds). When the switch machine performs a high-speed conversion operation, the system will conduct high-density sampling and in-depth analysis of the spectrum data within a few milliseconds before and after the operation to promptly detect sudden faults such as poor electrical contact during the instant and mechanical component collision impacts. At the same time, on the frequency scale, it deeply analyzes the relationship between signals in different frequency ranges and the special fault modes of high-speed switch machines. For example, for high-frequency vibration signals (above 2000 Hz), it focuses on checking the operating status of high-speed rotating components (such as motor rotors, gears, etc.). Once abnormal signal peaks or frequency offsets are found in the high-frequency band, further detailed detection is immediately carried out.

[0045] Extract fault feature parameters such as the peak value, mean value, variance, and frequency offset of the spectrum from the results of multi-scale spectrum analysis, and perform dimensionality reduction processing on these feature parameters to remove redundant information and highlight key features, so as to improve the efficiency and accuracy of fault diagnosis. Carefully compare the processed feature parameters with the pre-set strict fault thresholds. Since the tolerance for faults in high-speed railways is extremely low, the fault thresholds are set more strictly and precisely. If it is found that the mean value of the spectrum exceeds the normal range by 10%, the variance exceeds the normal range by 15%, the frequency offset exceeds the normal range by 10 Hz, and the spectrum peak exceeds the threshold by 30% in a specific frequency band (the frequency band related to key components), it is comprehensively judged that there is a potential fault in the switch machine. The system immediately issues a high-priority alarm signal to notify the operation and maintenance department of the high-speed railway to conduct an emergency inspection and handling. At the same time, through in-depth mining and analysis of the fault feature parameters, combined with the precise association model between the fault feature parameters and different fault types, quickly and accurately determine that the fault location is the loosening and oxidation of the contact point of a certain electrical connection component. After receiving the alarm, the operation and maintenance department quickly organizes a professional repair team, carries high-precision detection equipment and spare parts to the site. The repair team quickly locates the fault location according to the detailed fault information provided by the system, checks and repairs the electrical connection component, replaces the damaged contact point, and conducts a comprehensive electrical performance test to ensure that the switch machine resumes normal operation. After the repair is completed, the system records the repair process and results in detail, including the repair time, replaced components, test data after repair, etc., providing valuable reference for subsequent equipment maintenance and fault prevention.

[0046] The data storage module adopts a powerful distributed database storage system to meet the requirements of rapid storage and efficient access of a large amount of high-speed data in high-speed railways. Classify and store various types of collected data, including spectrum data, environmental parameters, adaptive spectrum analysis results, multi-scale spectrum analysis results, fault feature parameters, and fault diagnosis results, etc. in an orderly manner. To ensure the security and reliability of the data, a perfect data backup and recovery mechanism is established. Regularly perform full backups and incremental backups of the data. The backup data is stored in multiple different geographical locations and storage media respectively. At the same time, dynamically adjust the storage capacity allocation according to the importance and usage frequency of the data. Allocate more storage resources for recently frequently used data to improve the access speed, and perform reasonable compression and archiving processing on historical data to save storage space. In addition, regularly optimize and maintain the performance of the database, including index optimization, data cleaning, storage fragmentation reorganization, etc. operations to ensure the efficient operation of the database.

[0047] Through the application of this system, during the actual operation of high-speed railways, multiple potential faults in the switch machines have been discovered and handled in a timely manner, such as early faults in electrical components and minor damages to mechanical structures. This effectively ensures the safe and efficient operation of high-speed railways, fully demonstrating the important application value and technical advantages of this invention in the field of high-speed railways. At the same time, through in-depth analysis and mining of a large amount of historical data, the operation and maintenance departments of high-speed railways can summarize the potential laws and development trends of switch machine faults, formulate more scientific and reasonable maintenance plans and preventive measures in advance, further reducing the equipment failure rate and maintenance costs, and improving the overall operation efficiency of high-speed railways.

[0048] The above are only the preferred embodiments of the present invention, and do not impose any form of limitation on the present invention. Although the present invention has been disclosed above with the preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some changes or modifications to equivalent embodiments by using the disclosed technical content within the scope of the technical solution of the present invention. However, as long as it does not depart from the content of the technical solution of the present invention, any brief modifications, equivalent changes and modifications made to the above embodiments based on the technical essence of the present invention still fall within the scope of the technical solution of the present invention.

Claims

1. A fault diagnosis system for switch machine based on spectrum analysis, characterized in that: The system consists of the following components: Data acquisition module: Use high-precision acceleration, vibration, load, temperature and humidity sensors to collect spectrum data and environmental parameters of the switch machine in real time, and transmit the data through wired or wireless communication; Adaptive spectrum analysis module: Based on the load, temperature and humidity parameters transmitted by the data acquisition module, the algorithm that dynamically adjusts the frequency band range and resolution is used to analyze and process the collected spectrum data to capture the fault characteristic frequency; Multi-scale spectrum analysis module: divides the time scale into long and short time scales, divides the frequency into different ranges, analyzes the spectrum characteristics separately and fuses the results, judges the aging and wear of equipment through the long time scale, captures sudden failures through the short time scale, and analyzes fault information at different levels through the frequency scale, so as to fully grasp the operating status and fault conditions of the switch machine; Fault diagnosis and feature extraction module: Extracts the peak, mean, variance and frequency offset fault feature parameters of the spectrum from the multi-scale spectrum analysis results and fuses them, compares them with the preset threshold to judge the fault, determine the fault type and location, record and analyze the fault information, and provide data support for maintenance prevention; Data storage module: Use a database management system to allocate storage capacity based on data volume and storage ratio, classify and store various types of data, establish a backup and recovery mechanism, perform regular data management and maintenance, and ensure system operation and data availability.

2. A fault diagnosis system for switch machine based on spectrum analysis according to claim 1, characterized in that: The adaptive spectrum analysis module receives the load, temperature and humidity environmental parameters of the switch machine transmitted by the data acquisition module, monitors and analyzes these parameters in real time, establishes a correlation model between the environmental parameters and the spectrum analysis parameters, and dynamically adjusts the resolution and frequency band range parameters of the spectrum analysis according to the analysis results of the environmental parameters. When the load increases, the resolution of the spectrum analysis is adjusted to observe the spectrum details and capture the fault characteristic frequencies that may be caused by load changes. If the ambient temperature or humidity exceeds the normal range, the frequency band range is adjusted accordingly to cover the frequency range that changes due to environmental influences. The adjusted parameters are used to analyze and process the collected spectrum data. The spectrum analysis method is used to convert the vibration signal in the time domain into a frequency domain signal to obtain spectrum information. During the analysis process, the spectrum is pre-processed according to different application scenarios and requirements.

3. A fault diagnosis system for switch machine based on spectrum analysis according to claim 2, characterized in that: The adaptive spectrum analysis module dynamically adjusts the resolution and frequency band parameters of the spectrum analysis according to the analysis results of the environmental parameters. The resolution adjustment formula is: Among them, R new is the adjusted spectrum analysis resolution, R old is the spectrum analysis resolution before adjustment, β is the resolution adjustment coefficient, a constant determined experimentally and used to control the degree of resolution adjustment, n is the number of energy data points collected within a certain time window, and ΔE i is the difference between the switch machine operating energy and the average energy at the i-th time point, and the calculation formula is: Where E i is the energy value at the i-th time point, is the average energy value in the time window, E ref It is the reference energy value, which is the standard value preset according to the energy consumption of the switch machine during normal operation. The frequency band range parameter adjustment formula is: Among them, fmin new is the lower limit frequency of the adjusted frequency band, fmax new is the adjusted upper frequency of the frequency band, fmin old is the lower limit frequency of the frequency band before adjustment, fmax old is the upper frequency limit of the frequency band before adjustment. α1, α2, and α3 are the coefficients of the influence of load, temperature, and humidity on the frequency band range, respectively. They are obtained through a large number of experiments and data analysis and are used to quantify the effect of environmental factors on frequency band adjustment. ΔL is the difference between the current load and the standard load, ΔT is the difference between the current temperature and the standard temperature, and ΔH is the difference between the current humidity and the standard humidity.

4. A fault diagnosis system for switch machine based on spectrum analysis according to claim 3, characterized in that: The ΔL is the difference between the current load and the standard load, and its calculation formula is: ΔL = LL std , where L is the current load value, L std is the standard load value, ΔT is the difference between the current temperature and the standard temperature, and the calculation formula is: ΔT=TT std , where T is the current temperature value, T std is the standard temperature value, ΔH is the difference between the current humidity and the standard humidity, and the calculation formula is: ΔH=HH std , where H is the current humidity value, H std It is the standard humidity value.

5. The fault diagnosis system for switch machine based on spectrum analysis according to claim 1 is characterized in that: The multi-scale spectrum analysis module divides the time scale into long time scale and short time scale. The long time scale is used to analyze the macro trend of the switch machine operation. Under the long time scale, the spectrum data is statistically analyzed to calculate the mean and variance statistics of the spectrum. By observing the change of the statistics over time, the aging and wear trend of the equipment is judged. The short time scale is used to capture the instantaneous abnormal changes in the operation of the switch machine. Under the short time scale, the spectrum data is analyzed in high frequency details. The signal processing method is used to analyze the spectrum characteristics of each signal segment to find sudden fault events. On the frequency scale, the spectrum is divided into different frequency ranges, and the signal of each frequency range is analyzed separately. According to the structure and working principle of the switch machine, the potential connection between different frequency ranges and equipment failure is determined. By analyzing the signals of different frequency ranges, the spectrum characteristics that can reflect the fault information of the equipment at different levels are extracted. The analysis results of the time scale and frequency scale are integrated, and the spectrum characteristics under different time and frequency scales are comprehensively considered to more comprehensively understand the operation status and fault situation of the switch machine. At the same time, the spectrum characteristics of different frequency ranges are combined with the time scale analysis results to determine the specific location and type of the fault.

6. A fault diagnosis system for switch machine based on spectrum analysis according to claim 5, characterized in that: The multi-scale spectrum analysis module divides the time scale into long time scale and short time scale. The long time scale is used to analyze the macro trend of the switch machine operation. The analysis formula is: Among them, T trend It is a quantitative value of the change trend of the spectrum on a long time scale, which is used to evaluate the long-term state changes of equipment aging and wear. i is the spectral characteristic value at the i-th time point, S i+1 is the spectral eigenvalue at the i+1th time point, m is the total number of time points in the long time scale, Δt i is the i-th time interval, that is, the time difference between two adjacent time points.

7. The fault diagnosis system for switch machine based on spectrum analysis according to claim 5, characterized in that: The multi-scale spectrum analysis module divides the time scale into long time scale and short time scale. The short time scale is used to capture the instantaneous abnormal changes in the operation of the switch machine. The formula is: Among them, D detail It is a quantitative value of the prominence of high-frequency details in a short time scale, which is used to capture instantaneous abnormal changes. is the eigenvalue of the jth high-frequency component in the short time scale, is the eigenvalue of the kth high-frequency component in the short time scale, p is the number of high-frequency components in the short time scale, and w j is the weight coefficient of the jth high-frequency component, which is set according to the importance of the high-frequency component in fault diagnosis. The higher the importance, the greater the weight.

8. The fault diagnosis system for switch machine based on spectrum analysis according to claim 1, characterized in that: The fault diagnosis and feature extraction module extracts the peak value, mean value, variance and frequency offset fault feature parameters of the spectrum from the multi-scale spectrum analysis results and fuses them. The formula is: composite =γ1×F peak +γ2×F mean +γ3×F var +γ4×F offset , where F composite It is a comprehensive fault characteristic parameter, which is used to more comprehensively describe the fault status of the switch machine. peak It is the peak value of the spectrum, reflecting the amplitude of the frequency position where the energy in the spectrum is concentrated. mean is the mean of the spectrum, reflecting the average energy level of the spectrum, F var is the variance of the spectrum, which measures the discreteness of the spectrum energy distribution. offset is the frequency offset of the spectrum, indicating the deviation between the actual spectrum frequency and the standard frequency. γ1, γ2, γ3, and γ4 are the weight coefficients of peak value, mean value, variance, and frequency offset in the comprehensive fault feature parameters, respectively. They are obtained through the analysis of a large number of fault samples and the optimization of machine learning algorithms, and are used to determine the contribution of each feature parameter to fault diagnosis.

9. A method for diagnosing faults of a switch machine based on spectrum analysis, the method being applicable to a system for diagnosing faults of a switch machine based on spectrum analysis as claimed in any one of claims 1 to 8, characterized in that: The method comprises the following specific steps: Data acquisition: The data acquisition module is used to collect spectrum data and load, temperature and humidity parameters during the operation of the switch machine, and transmit the data; Adaptive spectrum analysis: Adaptive spectrum analysis module dynamically adjusts the resolution and frequency band parameters of spectrum analysis based on the collected environmental parameters using adaptive spectrum analysis algorithm, and performs adaptive spectrum analysis on the collected spectrum data; Multi-scale spectrum analysis: Use the multi-scale spectrum analysis module to analyze the spectrum data processed by adaptive spectrum analysis, and extract spectrum features at different time scales and frequency scales. In particular, analyze the spectrum change trend and calculate related statistics on a long time scale to determine the aging and wear of the equipment, and perform high-frequency detail analysis on a short time scale; Fault diagnosis and feature extraction: Based on the results of multi-scale spectrum analysis, the fault diagnosis and feature extraction module is used to extract fault feature parameters and perform dimensionality reduction processing. The processed fault feature parameters are compared with the pre-set fault threshold to determine whether there is a fault in the switch machine. If there is a fault, the fault type and location are determined, an alarm signal is issued, and the fault information is recorded in the data storage module; Data storage: The collected spectrum data, environmental parameters, adaptive spectrum analysis results, multi-scale spectrum analysis results, fault characteristic parameters and fault diagnosis result information are stored in the data storage module.

Images