A signal analysis method and system based on time domain spectrum

Abstract

The present disclosure relates to a time-domain map-based signal analysis method and system, comprising: collecting multi-channel vibration signals of a monitored device; preprocessing the channel vibration signals; performing time-domain mapping processing on the preprocessed multi-channel time-domain vibration signals to generate a time-domain map; extracting time-domain feature parameters from the time-domain map and performing frequency-domain transformation to obtain frequency-domain feature parameters; performing Hilbert transform demodulation processing on the preprocessed multi-channel time-domain vibration signals to obtain an envelope spectrum; calculating a real-time health index according to the time-domain feature parameters and the frequency-domain feature parameters; storing the real-time health index and the time-domain feature parameters in a time sequence database in chronological order, performing trend analysis, and obtaining a trend analysis result; based on the trend analysis result, combining a deep learning early warning model, and performing abnormal detection and pattern recognition on the device operating state; and visualizing the time-domain map, the envelope spectrum, the health index, and the early warning information in a graphical interface.

Description

Technical Field

[0001] The present disclosure relates to the technical field of power industrial equipment condition monitoring and fault diagnosis, and specifically relates to a signal analysis method and system based on time-domain maps, which is applicable to vibration signal analysis and operation state warning of rotating mechanical equipment. Background Art

[0002] In recent years, the installed capacity and power generation of China's power industry have continued to grow, power grid construction has been increasingly improved, and the scales of traditional energy and new energy power generation have expanded simultaneously. Power production has a significant impact on the national economy, and its safety must be maintained with long-term vigilance. Once a critical equipment fails, it will at least damage the equipment, and at worst lead to production shutdown or even serious accidents, causing huge economic losses. In industrial production, mechanical faults are widespread. Although different equipment has different fault types, they have similarities in characteristic quantities. Minor faults have limited impacts, but when the damage of rotating equipment intensifies, it may trigger cascading damage that spreads to the unit and production process, resulting in irreversible losses. It is necessary to collect operation data such as vibration, temperature and pressure before faults occur, so as to realize the identification and localization of potential or early faults.

[0003] Traditional vibration signal analysis methods mainly include two categories: time-domain analysis and frequency-domain analysis. Time-domain analysis describes the overall characteristics of signals through statistical characteristic parameters (such as root mean square value, peak value, kurtosis, etc.), but it is difficult to reveal the inherent frequency structure of signals; frequency-domain analysis converts signals to the frequency domain through Fast Fourier Transform (FFT), which can identify frequency components but loses time information, and cannot reflect the non-stationary characteristics of signals. In addition, the acquisition and analysis of multi-channel vibration signals in the prior art usually adopts the method of processing each independent channel separately, lacking the comprehensive analysis capability for the correlation between multi-channel signals, resulting in insufficiently comprehensive equipment condition assessment.

[0004] In recent years, with the development of Internet of Things technology and artificial intelligence technology, deep learning-based equipment condition monitoring methods have been widely applied. However, the existing fault diagnosis methods based on deep learning mainly have the following problems: (1) When the distribution of training data is inconsistent with that of test data, the model has poor generalization ability; (2) Alarms are only judged by relying on a single threshold, lacking effective identification capability for progressive anomalies such as the slow rising trend of vibration amplitude; (3) The fusion analysis capability for multi-source heterogeneous data (vibration, temperature, etc.) is insufficient, making it difficult to achieve accurate assessment of equipment health status.

[0005] Therefore, there is an urgent need for a signal analysis method and system that can comprehensively utilize the time-domain features of multi-channel vibration signals, realize the visual presentation of time-domain maps, and have intelligent trend warning capability. Contents of the Invention

[0006] To address the problems existing in the prior art, this disclosure proposes a signal analysis method and system based on time-domain spectra to solve at least one of the aforementioned technical problems. The technical solution adopted in this disclosure is as follows: In a first aspect, this disclosure provides a signal analysis method based on time-domain spectra, comprising the following steps: Multiple high-precision triaxial vibration sensors are arranged at different measuring points of the monitored equipment to synchronously collect multi-channel vibration signals of the monitored equipment at a preset sampling frequency. Each vibration sensor collects vibration acceleration signals, vibration velocity signals and vibration amplitude signals of the monitored equipment in the X-axis, Y-axis and Z-axis directions. The collected vibration signals from each channel are preprocessed to obtain preprocessed multi-channel time-domain vibration signals; The preprocessed multi-channel time-domain vibration signals are continuously mapped in the time domain according to the time sequence to generate a time-domain spectrum. The time-domain spectrum uses time as the horizontal axis and the vibration amplitude of each channel as the vertical axis, and the multi-channel vibration signals are superimposed and presented in the same time coordinate system. Time-domain feature parameters are extracted from the time-domain spectrum, and frequency-domain transformation is performed on the time-domain spectrum to obtain frequency-domain feature parameters, which include the amplitude information of the frequency shift and its harmonic components. The preprocessed multi-channel time-domain vibration signal is demodulated using Hilbert transform to extract the high-frequency resonance envelope reflecting the characteristics of the impact fault, generating an envelope signal. The envelope signal is then subjected to spectral analysis to obtain the envelope spectrum. Based on the time-domain feature parameters and the frequency-domain feature parameters, and combined with the preset equipment health evaluation model, the real-time health index of the monitored equipment is calculated. The real-time health index and the time-domain feature parameters are stored in a time-series database in chronological order to construct a historical data sequence. Trend analysis is then performed on the historical data sequence to obtain the trend analysis results. Based on the trend analysis results, combined with the deep learning early warning model, abnormal detection and pattern recognition are performed on the equipment operating status. If the upward slope of the vibration amplitude within the preset time window exceeds the preset trend threshold, or if the real-time health index is found to be lower than the preset health threshold, an early warning message is generated. The time-domain spectrum, the envelope spectrum, the health index, and the early warning information are presented visually using a graphical interface.

[0007] Optionally, the high-precision triaxial vibration sensor is a piezoelectric vibration sensor. The piezoelectric vibration sensor can be configured to achieve a measurement accuracy of ±0.5% and a sampling frequency of up to 20kHz.

[0008] Optionally, the sensor has a built-in embedded microprocessor that supports preprocessing operations such as filtering, integration, and fast Fourier transform on the acquired raw data.

[0009] Optionally, the preprocessing includes filtering, denoising, and integral transform processing.

[0010] Optionally, the time-domain mapping process includes: aligning the vibration signals of each channel according to the acquisition timestamp, mapping the signal amplitude of each channel to a visual curve based on a unified time axis, using different colors to identify the curves of different channels, and supporting graphical locking and depth measurement of any selected channel signal.

[0011] Optionally, the time-domain characteristic parameters include the effective value, peak value, peak-to-peak value, waveform factor, kurtosis factor, and margin factor of the vibration signal of each channel.

[0012] Optionally, the Hilbert transform demodulation process may include: The preprocessed multi-channel time-domain vibration signal is bandpass filtered to retain the signal components within the filtered frequency band, thus obtaining the filtered signal. Perform a Hilbert transform on the filtered signal to obtain an analytic signal; Calculate the amplitude of the analytic signal to obtain the envelope signal; The envelope signal is subjected to a fast Fourier transform to obtain the envelope spectrum, and the fault characteristic frequency and its harmonic components are extracted from the envelope spectrum.

[0013] Optionally, the equipment health assessment model adopts a vibration intensity assessment model based on the ISO 10816 series of standards.

[0014] Furthermore, the vibration intensity assessment model compares the time-domain characteristic parameters with the standard level range to determine the equipment operating status level.

[0015] Furthermore, the status level may include four levels: "good", "allowable", "poor" and "dangerous".

[0016] Optionally, the trend analysis includes: calculating the time series slope of the vibration amplitude within the preset time window; when the slope exceeds a preset proportion threshold, determining that the equipment is in a potential risk state; the preset trend threshold is dynamically adjusted by learning from historical operating data.

[0017] Optionally, the deep learning early warning model adopts a DANN-BiLSTM model based on a combination of a domain adaptive neural network and a bidirectional long short-term memory network.

[0018] Furthermore, the DANN-BiLSTM model includes: A feature extraction layer is used to extract high-dimensional feature representations from the time-domain feature parameters and the frequency-domain feature parameters; Domain adaptation layer is used to reduce the difference in feature distribution between the source domain and the target domain through adversarial training, thereby improving the model's generalization ability under different conditions. A bidirectional long short-term memory layer is used to capture the forward and backward dependencies of the temporal feature parameters in the time dimension; And a classification and early warning layer, which is used to output the device status classification results and the probability value of anomalies.

[0019] Optionally, the step of performing anomaly detection and pattern recognition on the equipment operating status based on the trend analysis results and combined with a deep learning early warning model, and generating early warning information when the upward slope of the vibration amplitude within the preset time window exceeds a preset trend threshold, or when the real-time health index is found to be lower than a preset health threshold, further includes: When potential fault symptoms are identified, the system calls upon the preset equipment knowledge graph and historical case library to perform root cause reasoning and severity assessment, generating an intelligent diagnostic report that includes fault location, development prediction, and maintenance recommendations.

[0020] Furthermore, the visualization includes: The health index statistics of the monitored equipment are presented in a pie chart; The historical changes in the health index of the monitored equipment are presented using a line graph; The real-time dynamic fluctuations of multi-channel vibration parameters are presented using a multi-curve overlay graph. Alarm statistics are presented in a bar chart. And real-time alarm information is presented in the form of a scrolling list.

[0021] A second aspect of this disclosure provides a signal analysis system based on time-domain spectra, the system comprising: The signal acquisition module is used to synchronously acquire multi-channel vibration signals of the monitored equipment at a preset sampling frequency by using multiple high-precision triaxial vibration sensors arranged at different measuring points of the monitored equipment. Each vibration sensor acquires vibration acceleration signals, vibration velocity signals and vibration amplitude signals of the monitored equipment in the X-axis, Y-axis and Z-axis directions. The signal preprocessing module is used to preprocess the acquired vibration signals from each channel to obtain preprocessed multi-channel time-domain vibration signals. The time-domain spectrum mapping module is used to perform continuous time-domain mapping processing on the preprocessed multi-channel time-domain vibration signals according to the time sequence to generate a time-domain spectrum; the time-domain spectrum uses time as the horizontal axis and the vibration amplitude of each channel as the vertical axis to superimpose and present the multi-channel vibration signals in the same time coordinate system. The feature extraction module is used to extract time-domain feature parameters from the time-domain spectrum, perform frequency domain transformation on the time-domain spectrum to obtain frequency-domain feature parameters, wherein the frequency-domain feature parameters include the amplitude information of the frequency shift and its harmonic components; The envelope analysis module is used to perform Hilbert transform demodulation on the preprocessed multi-channel time-domain vibration signal, extract the high-frequency resonance envelope reflecting the characteristics of the impact fault, generate the envelope signal, and perform spectral analysis on the envelope signal to obtain the envelope spectrum. The health index calculation module is used to calculate the real-time health index of the monitored equipment based on the time-domain feature parameters and the frequency-domain feature parameters, combined with a preset equipment health evaluation model. The trend analysis module is used to store the real-time health index and the time-domain feature parameters in a time-series database in chronological order, construct a historical data sequence, perform trend analysis on the historical data sequence, and obtain trend analysis results. The intelligent early warning module is used to detect anomalies and identify patterns in the equipment's operating status based on the trend analysis results and combined with a deep learning early warning model. If the upward slope of the vibration amplitude within the preset time window exceeds a preset trend threshold, or if the real-time health index is found to be lower than a preset health threshold, an early warning message is generated. The visualization module is used to visualize the time-domain spectrum, the envelope spectrum, the health index, and the early warning information using a graphical interface.

[0022] In a third aspect, this disclosure provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the signal analysis method based on time-domain spectra as described above.

[0023] In a fourth aspect, this disclosure provides an electronic device including a processor and a memory, the processor being configured to execute a computer program stored in the memory to implement the signal analysis method based on time-domain spectra as described above.

[0024] The beneficial effects of this disclosure are as follows: This disclosure provides a signal analysis method and system based on time-domain spectrograms. By using multi-channel vibration sensors, it collects multi-channel vibration signals from different parts and directions of the monitored equipment. Through comprehensive analysis of the multi-channel signals, it obtains more comprehensive equipment status information, thereby accurately assessing the health status of the equipment and effectively improving the system's ability to detect faults in different parts of the equipment, thus better supporting the comprehensive monitoring and maintenance of the equipment.

[0025] This disclosure utilizes Time-Domain Map (TDM) technology to perform real-time time-domain mapping and visualization of multi-channel vibration signals from equipment. By converting the vibration signals into visualized time-domain curves, users can intuitively observe the operating status of the equipment. This design enhances the system's visual monitoring capabilities, enabling equipment maintenance personnel to more quickly locate problems and take appropriate solutions.

[0026] This disclosure analyzes the historical data sequence to determine whether the vibration amplitude shows a continuous upward trend. Even if the vibration value does not exceed the set absolute threshold, it can be marked as a "potential risk" in advance, which can improve the system's ability to identify minor and potential faults. It can also provide early warning before the monitored equipment experiences obvious faults, thereby reducing the risk of equipment downtime.

[0027] Compared with the prior art, this disclosure has the following advantages: (1). This disclosure can realize the comprehensive visualization of vibration signals from different parts and directions of the monitored equipment through the synchronous acquisition of multi-channel vibration signals and time-domain spectrum mapping, which overcomes the information fragmentation problem caused by independent channel analysis in the prior art and improves the comprehensiveness and accuracy of equipment status assessment.

[0028] (2). This disclosure can combine time-domain graph technology with Hilbert envelope analysis, which can not only intuitively display the overall trend of equipment operation status through time-domain graph, but also extract periodic impact components hidden in complex background noise through envelope analysis, thus realizing early diagnosis of local defects such as bearing spalling and cracks.

[0029] (3). This disclosure can introduce an intelligent early warning function based on the DANN-BiLSTM model. It improves the generalization ability of the model under different working conditions through domain adaptation technology. It captures the forward and backward dependencies of the time series through a bidirectional long short-term memory network, and realizes the effective identification of gradual anomalies such as the slow upward trend of vibration amplitude. Even if the vibration value does not exceed the absolute threshold, it can be marked as a "potential risk" in advance, which significantly reduces the risk of sudden equipment failure.

[0030] (4). This disclosure can adopt a composite early warning strategy that combines vibration intensity assessment and trend analysis based on the ISO 10816 standard. The real-time operating status of the equipment is placed in a continuously evolving historical reference system for dynamic evaluation, realizing the transformation from "threshold alarm" to "trend early warning" and improving the timeliness and accuracy of the early warning.

[0031] (5) This disclosure can realize wireless data transmission through MESH networking, which has the characteristics of being decentralized, self-organizing, and multi-hop communication. It can automatically adjust the data transmission path, ensure the stability and reliability of data transmission in complex industrial environments, and reduce the cabling cost of system deployment. Attached Figure Description

[0032] The accompanying drawings, which form part of this application, are used to provide a further understanding of this disclosure. The illustrative embodiments of this disclosure and their descriptions are used to explain this disclosure and do not constitute an undue limitation of this disclosure.

[0033] Figure 1 This is a flowchart of the signal analysis method based on time-domain spectra as described in Embodiment 1 of this disclosure.

[0034] Figure 2 This is an architecture diagram of the signal analysis system based on time-domain spectrograms as described in Embodiment 2 of this disclosure. Detailed Implementation

[0035] The present disclosure will now be described in detail with reference to the accompanying drawings and embodiments. It should be noted that, unless otherwise specified, the embodiments and features described in the present application can be combined with each other.

[0036] The following detailed descriptions are exemplary and intended to provide further detailed explanation of this disclosure. Unless otherwise specified, all technical terms used in this disclosure have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains. The terminology used in this disclosure is for the purpose of describing particular embodiments only and is not intended to limit the exemplary embodiments according to this disclosure.

[0037] Example 1: like Figure 1 As shown, this embodiment provides a signal analysis method based on time-domain spectra, applied to the monitoring of the operating status and fault early warning of important auxiliary pumps in power plants (including stator cooling water pumps, lubricating oil pumps, and EH oil pumps). The method includes the following steps: S100, Acquire multi-channel vibration signals: A high-precision triaxial vibration sensor is arranged at each of the front and rear ends of the pump body being monitored. Each sensor is fixed to the surface of the pump body by magnetic attraction. The triaxial vibration sensor is a piezoelectric triaxial vibration sensor, and its sensitive element is made of piezoelectric material. When the piezoelectric material is subjected to force, a charge is generated on its surface. This charge is amplified by a charge amplifier and a measuring circuit, and after impedance transformation, it becomes an electrical output proportional to the applied external force.

[0038] The triaxial vibration sensor features wide bandwidth, high sensitivity, high signal-to-noise ratio, simple structure, reliable operation, and light weight. It can measure the vibration acceleration, vibration velocity, and vibration amplitude signals of the pump body under test in the X, Y, and Z axes. The vibration acceleration, vibration velocity, and vibration amplitude signals are used as the acquired multi-channel vibration signals.

[0039] The triaxial vibration sensor achieves a measurement accuracy of ±0.5% and a sampling frequency of up to 20kHz. The sensor incorporates an embedded microprocessor, supporting intelligent data preprocessing functions. It can perform filtering, integration, and FFT transformation on the acquired raw data to reduce data redundancy and improve transmission efficiency. Furthermore, the triaxial vibration sensor has a self-diagnostic function, promptly sending alarm messages to the system when abnormal signals or hardware failures occur.

[0040] S200, Signal Preprocessing: The acquired multi-channel vibration signals were first bandpass filtered to remove high-frequency noise and low-frequency drift interference. Then, an integral transform was performed to convert the acceleration signals into velocity signals, facilitating subsequent comparative analysis with the ISO 10816 standard. After preprocessing, the preprocessed multi-channel vibration signals were obtained.

[0041] S300, Time-Domain Spectrum Mapping: The preprocessed multi-channel vibration signals are aligned according to the acquisition timestamps, and the amplitude of each channel signal is mapped to a visual curve using a unified time axis as a reference. The time-domain graph uses time as the horizontal axis and the vibration amplitude of each channel as the vertical axis, superimposing the multi-channel vibration signals in the same time coordinate system. Different channels are indicated by different colors; for example, the X-axis signal is represented by a red curve, the Y-axis signal by a green curve, and the Z-axis signal by a blue curve. Users can lock onto and measure the depth of any selected channel signal through a graphical interface, including functions such as cursor measurement and waveform zooming.

[0042] S400, Feature Extraction: Time-domain characteristic parameters can be extracted from the time-domain spectrum, including: RMS, peak value, peak-to-peak value, form factor, kurtosis factor, and crest factor. Simultaneously, a Fast Fourier Transform (FFT) is performed on the vibration signals of each channel to extract frequency-domain characteristic parameters, including the rotational frequency (1X) and the amplitude information of its harmonic components (2X, 3X, etc.). By analyzing the amplitude evolution of the rotational frequency and its harmonic components, the spectral characteristic changes caused by typical fault modes such as mass imbalance (prominent 1X component) or shaft misalignment (significant 2X component) can be identified.

[0043] S500, Envelope Analysis: The preprocessed vibration signal is demodulated using Hilbert transform. First, the original vibration signal is bandpass filtered to retain signal components within the filtered frequency band. Then, a Hilbert transform is performed on the filtered signal to obtain an analytic signal. The amplitude of the analytic signal is calculated to obtain the envelope signal. Finally, a fast Fourier transform is performed on the envelope signal to obtain the envelope spectrum. The envelope spectrum effectively separates periodic impact components hidden in complex background noise, accurately locating early damage to specific bearing components and greatly improving the early diagnosis capability of local defects such as bearing spalling and cracks.

[0044] S600 Health Index Calculation: Based on the time-domain and frequency-domain characteristic parameters, and combined with the equipment health assessment model based on the ISO 10816 series of standards, the real-time health index of the monitored equipment is calculated. The vibration intensity map transforms abstract spectral data into a quantitative assessment of equipment status levels conforming to the ISO 10816 standard. By comparing the current measurement value with a level range based on the statistical distribution of historical data, the equipment operating status level is determined. The status levels include four levels: "Good," "Allowable," "Poor," and "Dangerous."

[0045] S700, Trend Analysis: The real-time health index and the time-domain feature parameters are stored in a time-series database in chronological order to construct a historical data sequence. Trend analysis is performed on the historical data sequence to calculate the time-series slope of the vibration amplitude within a preset time window (e.g., 30 minutes). When the vibration amplitude is detected to be continuously rising within the preset time window, and the rising slope exceeds a preset proportional threshold (e.g., rising more than 5% per minute), the device is determined to be in a "potential risk" state, even if the current vibration value has not reached the absolute threshold. The preset trend threshold is dynamically adjusted through learning from historical operating data to adapt to the operating characteristics of different devices.

[0046] S800, Intelligent Early Warning: Based on the trend analysis results, and combined with the DANN-BiLSTM deep learning early warning model, anomaly detection and pattern recognition are performed on the equipment operating status. The DANN-BiLSTM model includes a feature extraction layer, a domain adaptation layer, a bidirectional long short-term memory layer, and a classification early warning layer. The feature extraction layer extracts high-dimensional feature representations from the time-domain and frequency-domain feature parameters. The domain adaptation layer introduces a gradient reversal layer through adversarial training, reversing the gradient direction during backpropagation, so that the feature representations learned by the feature extractor have a similar distribution between the source and target domains, thereby improving the model's generalization ability under different operating conditions. The bidirectional long short-term memory layer captures the forward and backward dependencies of the time series through forward LSTM and backward LSTM respectively, fully exploring the dynamic patterns in the time-series data. The classification early warning layer outputs the equipment status classification results (normal, potential risk, abnormal, dangerous) and the corresponding anomaly probability values.

[0047] When potential fault symptoms are identified, the system invokes the embedded equipment knowledge graph and historical case library to perform root cause reasoning and severity assessment. The equipment knowledge graph includes equipment structural information, a fault mode library, a fault feature library, and a maintenance strategy library. It uses a graph neural network (GNN) for knowledge reasoning to generate an intelligent diagnostic report that includes fault location, development prediction, and maintenance recommendations. Simultaneously, the system continuously learns and incorporates these diagnostic cases into new knowledge, enabling the self-evolution and optimization of its diagnostic capabilities.

[0048] S900, Visual Presentation: The analysis results are presented graphically, including: a pie chart showing equipment health index statistics; a line chart showing the historical changes in equipment health index; a multi-curve overlay chart showing real-time dynamic fluctuations of multi-channel vibration parameters; a bar chart showing monthly alarm statistics; and a scrolling list showing real-time alarm information. Users can access the system via PC, mobile phone, or tablet for remote monitoring and management.

[0049] Example 2: like Figure 2 As shown, in a second aspect, this disclosure provides a signal analysis system based on time-domain spectra, the system comprising: The signal acquisition module 100 is used to synchronously acquire multi-channel vibration signals of the monitored equipment at a preset sampling frequency by using multiple high-precision triaxial vibration sensors arranged at different measuring points of the monitored equipment. Each vibration sensor acquires vibration acceleration signals, vibration velocity signals and vibration amplitude signals of the monitored equipment in the X-axis, Y-axis and Z-axis directions. The signal preprocessing module 200 is used to preprocess the acquired vibration signals from each channel to obtain preprocessed multi-channel time-domain vibration signals. The time-domain spectrum mapping module 300 is used to perform continuous time-domain mapping processing on the preprocessed multi-channel time-domain vibration signal according to the time sequence to generate a time-domain spectrum; the time-domain spectrum uses time as the horizontal axis and the vibration amplitude of each channel as the vertical axis to superimpose and present the multi-channel vibration signal in the same time coordinate system. The feature extraction module 400 is used to extract time-domain feature parameters from the time-domain spectrum, perform frequency domain transformation on the time-domain spectrum to obtain frequency-domain feature parameters, wherein the frequency-domain feature parameters include the amplitude information of the frequency shift and its harmonic components; The envelope analysis module 500 is used to perform Hilbert transform demodulation processing on the preprocessed multi-channel time-domain vibration signal, extract the high-frequency resonance envelope reflecting the characteristics of the impact fault, generate the envelope signal, and perform spectral analysis on the envelope signal to obtain the envelope spectrum. The health index calculation module 600 is used to calculate the real-time health index of the monitored equipment based on the time-domain feature parameters and the frequency-domain feature parameters, combined with a preset equipment health evaluation model. The trend analysis module 700 is used to store the real-time health index and the time-domain feature parameters in a time-series database in chronological order, construct a historical data sequence, perform trend analysis on the historical data sequence, and obtain trend analysis results. The intelligent early warning module 800 is used to detect abnormalities and identify patterns in the operating status of the equipment based on the trend analysis results and combined with the deep learning early warning model. If the upward slope of the vibration amplitude within the preset time window exceeds the preset trend threshold, or if the real-time health index is found to be lower than the preset health threshold, an early warning message is generated. The visualization module 900 is used to visualize the time-domain spectrum, the envelope spectrum, the health index, and the early warning information using a graphical interface.

[0050] It should be noted that the system described in Embodiment 2 is only one system implementation of the method described in Embodiment 1, and does not mean that the method described in Embodiment 1 necessarily depends on the system described in Embodiment 2.

[0051] Example 3: Embodiment 3 of this disclosure provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, it implements the signal analysis method based on time-domain spectra as described in Embodiment 1.

[0052] The computer-readable storage medium includes volatile or non-volatile, removable or non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, computer program modules, or other data). Computer-readable storage media include, but are not limited to, RAM (Random Access Memory), ROM (Read-Only Memory), EEPROM (Electrically Erasable Programmable Read-Only Memory), flash memory or other memory technologies, CD-ROM (Compact Disc Read-Only Memory), DVD or other optical disc storage, cartridges, magnetic tapes, disk storage or other magnetic storage devices, or any other medium that can be used to store desired information and is accessible to a computer.

[0053] Example 4: Embodiment 4 of this disclosure provides an electronic device, which includes a processor and a memory, wherein the processor is used to execute a computer program stored in the memory to implement the signal analysis method based on time-domain spectra described in Embodiment 1.

[0054] Based on this understanding, all or part of the processes in the methods of the above embodiments can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, and read-only memory (ROM).

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

[0056] This disclosure is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this disclosure. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create a machine for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0057] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0058] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0059] In the description of this specification, references to terms such as "an embodiment," "example," "specific example," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this disclosure. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0060] In summary, the signal analysis method and system based on time-domain graphs provided in embodiments 1-4 of this disclosure acquire multi-channel vibration signals from different parts and directions of the monitored equipment using multi-channel vibration sensors. Through comprehensive analysis of these multi-channel signals, more complete equipment status information is obtained, thereby accurately assessing the equipment's health condition and effectively improving the system's ability to detect faults in different parts of the equipment, thus better supporting comprehensive equipment monitoring and maintenance. This disclosure utilizes time-domain graph (TDM) technology to perform real-time time-domain mapping and visualization of the equipment's multi-channel vibration signals. By converting the vibration signals into visualized time-domain curves, users can intuitively observe the equipment's operating status. This design enhances the system's visual monitoring capabilities, enabling equipment maintenance personnel to more quickly locate problems and take corresponding solutions.

[0061] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this disclosure and not to limit them. Although this disclosure has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of this disclosure. Any modifications or equivalent substitutions that do not depart from the spirit and scope of this disclosure should be covered within the protection scope of the claims of this disclosure.

Claims

1. A signal analysis method based on time-domain spectra, characterized in that, Includes the following steps: Multiple high-precision triaxial vibration sensors are arranged at different measuring points of the monitored equipment to synchronously collect multi-channel vibration signals of the monitored equipment at a preset sampling frequency. Each vibration sensor collects vibration acceleration signals, vibration velocity signals and vibration amplitude signals of the monitored equipment in the X-axis, Y-axis and Z-axis directions. The collected vibration signals from each channel are preprocessed to obtain preprocessed multi-channel time-domain vibration signals; The preprocessed multi-channel time-domain vibration signals are continuously mapped in the time domain according to the time sequence to generate a time-domain spectrum. The time-domain spectrum uses time as the horizontal axis and the vibration amplitude of each channel as the vertical axis, and the multi-channel vibration signals are superimposed and presented in the same time coordinate system. Time-domain feature parameters are extracted from the time-domain spectrum, and frequency-domain transformation is performed on the time-domain spectrum to obtain frequency-domain feature parameters, which include the amplitude information of the frequency shift and its harmonic components. The preprocessed multi-channel time-domain vibration signal is demodulated using Hilbert transform to extract the high-frequency resonance envelope reflecting the characteristics of the impact fault, generating an envelope signal. The envelope signal is then subjected to spectral analysis to obtain the envelope spectrum. Based on the time-domain feature parameters and the frequency-domain feature parameters, and combined with the preset equipment health evaluation model, the real-time health index of the monitored equipment is calculated. The real-time health index and the time-domain feature parameters are stored in a time-series database in chronological order to construct a historical data sequence. Trend analysis is then performed on the historical data sequence to obtain the trend analysis results. Based on the trend analysis results, combined with the deep learning early warning model, abnormal detection and pattern recognition are performed on the equipment operating status. If the upward slope of the vibration amplitude within the preset time window exceeds the preset trend threshold, or if the real-time health index is found to be lower than the preset health threshold, an early warning message is generated. The time-domain spectrum, the envelope spectrum, the health index, and the early warning information are presented visually using a graphical interface.

2. The method according to claim 1, characterized in that, The high-precision triaxial vibration sensor is a piezoelectric vibration sensor.

3. The method according to claim 1, characterized in that, The time-domain mapping process includes: aligning the vibration signals of each channel according to the acquisition timestamp, mapping the signal amplitude of each channel to a visual curve based on a unified time axis, using different colors to identify the curves of different channels, and supporting graphical locking and depth measurement of any selected channel signal.

4. The method according to claim 1, characterized in that, The Hilbert transform demodulation process includes: The preprocessed multi-channel time-domain vibration signal is bandpass filtered to retain the signal components within the filtered frequency band, thus obtaining the filtered signal. Perform a Hilbert transform on the filtered signal to obtain an analytic signal; Calculate the amplitude of the analytic signal to obtain the envelope signal; The envelope signal is subjected to a fast Fourier transform to obtain the envelope spectrum, and the fault characteristic frequency and its harmonic components are extracted from the envelope spectrum.

5. The method according to claim 1, characterized in that, The equipment health assessment model adopts a vibration intensity assessment model based on the ISO10816 series of standards.

6. The method according to claim 1, characterized in that, The vibration intensity assessment model compares the time-domain characteristic parameters with the standard level range to determine the equipment operating status level.

7. The method according to claim 1, characterized in that, The trend analysis includes: calculating the time series slope of the vibration amplitude within the preset time window; when the slope exceeds a preset ratio threshold, the equipment is determined to be in a potential risk state; the preset trend threshold is dynamically adjusted by learning from historical operating data.

8. The method according to claim 1, characterized in that, In step S800, the deep learning early warning model adopts the DANN-BiLSTM model, which combines a domain adaptive neural network and a bidirectional long short-term memory network.

9. The method according to claim 1, characterized in that, The method of generating an early warning message based on the trend analysis results and combined with a deep learning early warning model to detect anomalies and identify patterns in the equipment's operating status, and which generates an early warning message when the upward slope of the vibration amplitude within the preset time window exceeds a preset trend threshold, or when the real-time health index is found to be lower than a preset health threshold, also includes: When potential fault symptoms are identified, the system calls upon the preset equipment knowledge graph and historical case library to perform root cause reasoning and severity assessment, generating an intelligent diagnostic report that includes fault location, development prediction, and maintenance recommendations.

10. A signal analysis system based on time-domain spectra, characterized in that, include: The signal acquisition module is used to synchronously acquire multi-channel vibration signals of the monitored equipment at a preset sampling frequency by using multiple high-precision triaxial vibration sensors arranged at different measuring points of the monitored equipment. Each vibration sensor acquires vibration acceleration signals, vibration velocity signals and vibration amplitude signals of the monitored equipment in the X-axis, Y-axis and Z-axis directions. The signal preprocessing module is used to preprocess the acquired vibration signals from each channel to obtain preprocessed multi-channel time-domain vibration signals. The time-domain spectrum mapping module is used to perform continuous time-domain mapping processing on the preprocessed multi-channel time-domain vibration signals according to the time sequence to generate a time-domain spectrum; the time-domain spectrum uses time as the horizontal axis and the vibration amplitude of each channel as the vertical axis to superimpose and present the multi-channel vibration signals in the same time coordinate system. The feature extraction module is used to extract time-domain feature parameters from the time-domain spectrum, perform frequency domain transformation on the time-domain spectrum to obtain frequency-domain feature parameters, wherein the frequency-domain feature parameters include the amplitude information of the frequency shift and its harmonic components; The envelope analysis module is used to perform Hilbert transform demodulation on the preprocessed multi-channel time-domain vibration signal, extract the high-frequency resonance envelope reflecting the characteristics of the impact fault, generate the envelope signal, and perform spectral analysis on the envelope signal to obtain the envelope spectrum. The health index calculation module is used to calculate the real-time health index of the monitored equipment based on the time-domain feature parameters and the frequency-domain feature parameters, combined with a preset equipment health evaluation model. The trend analysis module is used to store the real-time health index and the time-domain feature parameters in a time-series database in chronological order, construct a historical data sequence, perform trend analysis on the historical data sequence, and obtain trend analysis results. The intelligent early warning module is used to detect anomalies and identify patterns in the equipment's operating status based on the trend analysis results and combined with a deep learning early warning model. If the upward slope of the vibration amplitude within the preset time window exceeds a preset trend threshold, or if the real-time health index is found to be lower than a preset health threshold, an early warning message is generated. The visualization module is used to visualize the time-domain spectrum, the envelope spectrum, the health index, and the early warning information using a graphical interface.

Images