Online monitoring and early warning system, method, equipment and medium for generator carbon brush condition

Abstract

This invention discloses an online monitoring and early warning system, method, device, and medium for generator carbon brush condition, relating to the field of generator carbon brush condition monitoring technology. The system includes: acquiring operating status data of each carbon brush on the same polarity brush holder of a generator; constructing an equivalent circuit model of the carbon brush based on the operating status data to obtain the electrical contact state parameters of each carbon brush; constructing a thermal-electric decoupling model based on the operating status data to separate and obtain the pure mechanical frictional heat parameters of each carbon brush; constructing a time-frequency joint analysis model based on the operating status data to extract the micro-arc discharge parameters of each carbon brush; fusing the electrical contact state parameters, pure mechanical frictional heat parameters, and micro-arc discharge parameters into a model to obtain a comprehensive health index for each carbon brush, and outputting an early warning command based on the comprehensive health index. This invention not only achieves precise decoupling of multi-physics field characteristics but also significantly improves the accuracy of carbon brush fault early warning.

Description

Technical Field

[0001] This invention relates to the field of generator carbon brush condition monitoring technology, and more specifically, to a generator carbon brush condition online monitoring and early warning system, method, device and medium. Background Technology

[0002] Generator carbon brushes are the core conductive components connecting rotating slip rings and stationary circuits. Real-time monitoring of carbon brush operating status and accurate early warning are of significant engineering importance for ensuring the safe and stable operation of generator sets and reliable power supply from the power grid. Currently, existing monitoring technologies typically rely on comparisons of single thresholds for alarms, which cannot eliminate complex environmental interference. Furthermore, due to the lack of effective decoupling of heat sources, it is not only difficult to accurately distinguish between mechanical jamming and normal electrical heating, but it is also prone to missing hidden micro-arcs, resulting in poor monitoring reliability. Summary of the Invention

[0003] The purpose of this invention is to provide an online monitoring and early warning system, method, device and medium for generator carbon brush status, which not only achieves precise decoupling of multi-physics field characteristics, but also significantly improves the accuracy of carbon brush fault early warning.

[0004] This invention is achieved through the following technical solution: An online monitoring and early warning system for generator carbon brush condition includes: The data acquisition module is used to acquire the operating status data of each carbon brush on the same polarity brush holder of the generator. The operating status data includes shunt current data, temperature data, and high-frequency characteristic data. The contact state modeling module, connected to the data acquisition module, is used to construct an equivalent circuit model of the carbon brush based on the shunt current data, and obtain the electrical contact state parameters of each carbon brush. The heat source separation modeling module is connected to the data acquisition module and is used to construct a thermal-electric decoupling model based on the temperature data and shunt current data to separate and obtain the pure mechanical frictional heat parameters of each carbon brush. The discharge characteristic modeling module is connected to the data acquisition module and is used to construct a time-frequency joint analysis model based on the high-frequency characteristic data to extract the micro-arc discharge parameters of each carbon brush. The state fusion and early warning module is connected to the contact state modeling module, the heat source separation modeling module, and the discharge characteristic modeling module, respectively. It is used to fuse the electrical contact state parameters, pure mechanical friction heat parameters, and micro-arc discharge parameters into a model to obtain the comprehensive health index of each carbon brush, and output an early warning command based on the comprehensive health index.

[0005] Specifically, the contact state modeling module performs the following steps: The generator slip ring and the parallel carbon brushes on the same polarity brush holder are used to construct the equivalent circuit model of the carbon brush. The shunt current data of each carbon brush is used as an input variable and substituted into the equivalent circuit model of the carbon brush to solve for impedance, thereby obtaining the real-time equivalent contact resistance of each carbon brush, and the real-time equivalent contact resistance is used as the electrical contact state parameter.

[0006] Specifically, the heat source separation modeling module executes the following steps: An electrothermal calculation model is constructed based on the shunt current data of each carbon brush to calculate the electrothermal component caused by the current in each carbon brush. Extract the background ambient temperature inside the generator; Subtract the corresponding electrothermal component and background ambient temperature from the temperature data of each carbon brush to separate the heat generated by mechanical friction between the carbon brush and the slip ring, and use it as the pure mechanical frictional heat parameter.

[0007] Specifically, the discharge feature modeling module performs the following steps: The high-frequency characteristic data includes high-frequency current signals and wideband vibration signals; The energy envelopes of the high-frequency current signal and the wideband vibration signal within a set high-frequency band are extracted respectively. Calculate the cross-correlation coefficient between the energy envelope of a high-frequency current signal and the energy envelope of a broadband vibration signal; When the cross-correlation coefficient is greater than a preset threshold, a real micro-arc discharge is confirmed, and the energy integral value of the high-frequency current signal is calculated and used as the micro-arc discharge parameter.

[0008] Specifically, the state fusion and early warning module performs the following steps: A multidimensional state-space model is constructed with electrical contact state parameters, pure mechanical frictional heat parameters, and micro-arc discharge parameters as dimensions. The real-time parameters of each carbon brush are mapped to a multi-dimensional state space model to form real-time state points; Calculate the spatial distance between the real-time status point and the baseline health point in the multidimensional state space model, and determine the comprehensive health index of each carbon brush based on the magnitude of the spatial distance.

[0009] Specifically, the state fusion and early warning module further includes the following execution steps: When the electrical contact state parameters are detected to be gradually increasing and the pure mechanical friction heat parameters are within the set normal range, the carbon brush is determined to be in a normal state of spring pressure decay. When a step change is detected in the electrical contact status parameters and the pure mechanical frictional heat parameters exceed the set normal range, the carbon brush is determined to be stuck or the spring is broken. When the micro-arc discharge parameters are continuously non-zero and show an upward trend, the carbon brush is determined to be in an electro-ablation state. Specifically, it also includes a lifespan prediction module, which is used for: Obtain the historical cumulative values ​​of pure mechanical frictional heat parameters and convert them into mechanical wear. The historical cumulative values ​​of micro-arc discharge parameters are obtained and converted into electrical ablation wear. By substituting the mechanical wear and electrical ablation wear into a preset wear accumulation model, the remaining time before the comprehensive health index reaches the scrap threshold is predicted, and the remaining service life of the carbon brush is output.

[0010] A method for online monitoring and early warning of generator carbon brush condition includes: The operating status data of each carbon brush on the same polarity brush holder of the generator is obtained, and the operating status data includes shunt current data, temperature data and high frequency characteristic data. Based on the shunt current data, an equivalent circuit model of the carbon brush is constructed to obtain the electrical contact state parameters of each carbon brush. Based on the temperature data and shunt current data, a thermal-electric decoupling model is constructed to separate the pure mechanical frictional heat parameters of each carbon brush. Based on the high-frequency characteristic data, a time-frequency joint analysis model is constructed to extract the micro-arc discharge parameters of each carbon brush. The electrical contact state parameters, pure mechanical frictional heat parameters, and micro-arc discharge parameters are fused into a model to obtain a comprehensive health index for each carbon brush, and an early warning command is output based on the comprehensive health index.

[0011] An electronic device, comprising: At least one processor; and A memory that is communicatively connected to the at least one processor; The memory stores a computer program that can be executed by the at least one processor, which enables the at least one processor to perform the functions of the generator carbon brush condition online monitoring and early warning system.

[0012] A computer-readable storage medium storing computer instructions for instructing a computer to perform the functions of an online monitoring and early warning system for generator carbon brush status.

[0013] The technical solution of the present invention has at least the following advantages and beneficial effects: This invention separates pure mechanical frictional heat by constructing a thermo-electric decoupling model and extracts micro-arc characteristics by combining time-frequency analysis. This not only completely eliminates background interference caused by fluctuations in operating conditions, but also achieves multi-dimensional and accurate assessment of carbon brush health, greatly improving the early detection rate of potential problems. Attached Figure Description

[0014] Figure 1 Schematic diagram of the generator carbon brush condition online monitoring and early warning system provided by the present invention; Figure 2 The flowchart of the generator carbon brush condition online monitoring and early warning method provided by the present invention. Detailed Implementation

[0015] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0016] Example 1 like Figure 1 As shown, this embodiment provides an online monitoring and early warning system for generator carbon brush status. This system constructs a complete closed-loop diagnostic architecture from bottom-level data acquisition to top-level strategy output. The system mainly includes a data acquisition module, a contact state modeling module, a heat source separation modeling module, a discharge characteristic modeling module, and a status fusion and early warning module. These modules are sequentially connected via preset logical instructions and data flow channels to collaboratively complete in-depth analysis and multi-dimensional early warning of the generator carbon brush operating status.

[0017] To address the complex electromagnetic and thermodynamic environment inside the generator, the data acquisition module aims to obtain operational status data for each carbon brush on the same polarity brush holder. This operational status data includes shunt current data, temperature data, and high-frequency characteristic data. In a preferred implementation scenario, multiple carbon brushes are typically connected in parallel on the same polarity brush holder. The current-carrying distribution and mechanical contact state of each carbon brush are closely related. By acquiring multi-dimensional operational status data, the macroscopic and microscopic operating conditions of the carbon brushes can be reflected from different physical field perspectives.

[0018] Specifically, in the data acquisition process, the data acquisition module employs differentiated sampling logic based on the frequency band characteristics of different physical signals. For shunt current and temperature data, the system uses a low-frequency sampling channel for timed acquisition, primarily to capture quasi-static trend parameters that change slowly with generator load and environment. For high-frequency characteristic data, the system uses a high-frequency sampling channel for event-triggered acquisition to capture transient anomaly signals within a very short time window. To ensure strict temporal correspondence of cross-physical field data in subsequent multi-dimensional fusion analysis, the system timestamps the data from the low-frequency and high-frequency sampling channels using the system bus clock, generating operational status data with a unified time base.

[0019] The contact state modeling module is connected to the data acquisition module to construct an equivalent circuit model of the carbon brushes based on the shunt current data, thereby obtaining the electrical contact state parameters of each carbon brush. Specifically, this module constructs an equivalent circuit model of the generator slip ring and the parallel carbon brushes on the same polarity brush holder. In the equivalent circuit model, the generator slip ring body serves as the common reference terminal for current collection, the busbar of the same polarity brush holder serves as the common injection terminal, and each parallel carbon brush and its interface with the slip ring constitute multiple parallel time-varying impedance branches. Subsequently, the shunt current data of each carbon brush is substituted into the equivalent circuit model as an input variable for impedance calculation. Based on the dynamic distribution ratio of the current in each branch, the real-time equivalent contact resistance of each branch is analyzed in reverse, and the real-time equivalent contact resistance is used as an electrical contact state parameter. The variation law of the real-time equivalent contact resistance directly and objectively reflects the electrical conduction quality of the interface between the bottom surface of the carbon brush and the slip ring.

[0020] The heat source separation modeling module is connected to the data acquisition module, aiming to solve the technical challenge of accurately defining the cause of carbon brush heating under complex operating conditions. Based on temperature and shunt current data, this module constructs a thermo-electric decoupling model to separate the pure mechanical frictional heat parameters of each carbon brush. At the execution level, firstly, an electrothermal calculation model is built based on the shunt current data of each carbon brush. According to the Joule heating effect, the electrothermal component caused solely by the current passing through the brush is calculated. Simultaneously, the background ambient temperature inside the generator is extracted to eliminate common-mode thermal interference caused by seasonal changes and cooling air temperature fluctuations. Finally, the corresponding electrothermal component and background ambient temperature are subtracted from the real-time temperature data of each carbon brush. This mathematically isolates electromagnetic heating and environmental influences, separating the heat generated purely by mechanical friction during high-speed relative sliding between the carbon brush and the slip ring, and uses this as the pure mechanical frictional heat parameter. The extraction of the pure mechanical frictional heat parameter provides a clean physical basis for accurately assessing the mechanical jamming state and spring clamping force state inside the brush holder.

[0021] The discharge characteristic modeling module is connected to the data acquisition module to construct a time-frequency joint analysis model based on high-frequency characteristic data, extracting the micro-arc discharge parameters of each carbon brush. The high-frequency characteristic data includes not only high-frequency current signals but also broadband vibration signals. When a micro-arc discharge occurs, it not only excites high-frequency oscillating current in the electrical circuit but also triggers a microscopic acoustic explosion impact at the contact interface. Based on this physical mechanism, the discharge characteristic modeling module extracts the energy envelopes of the high-frequency current signal and the broadband vibration signal within a set high-frequency band. Next, it calculates the cross-correlation coefficient between the energy envelopes of the high-frequency current signal and the broadband vibration signal. The calculation of the cross-correlation coefficient is intended for cross-validation; if and only if the cross-correlation coefficient is greater than a preset threshold, it is confirmed that the currently captured anomaly is not electromagnetic noise interference from the generator's external space or a simple mechanical impact, but rather a genuine micro-arc discharge phenomenon. After confirming the occurrence of a genuine discharge, the energy envelope of the high-frequency current signal is integrated to evaluate the total energy released in a single or continuous discharge, and this is used as the micro-arc discharge parameter.

[0022] The state fusion and early warning module is connected to the contact state modeling module, the heat source separation modeling module, and the discharge characteristic modeling module, respectively, to perform top-level aggregation of multi-source data. This module fuses electrical contact state parameters, pure mechanical frictional heat parameters, and micro-arc discharge parameters into a model to obtain a comprehensive health index for each carbon brush, and outputs early warning commands based on the comprehensive health index.

[0023] Specifically, in terms of the computational logic of state fusion, the state fusion and early warning module constructs a multi-dimensional state-space model with electrical contact state parameters, pure mechanical frictional heat parameters, and micro-arc discharge parameters as independent dimensions. The real-time parameters of each carbon brush are mapped into the multi-dimensional state-space model, forming real-time state points reflecting the current operating state of the carbon brushes. Simultaneously, historical baseline data from the initial healthy operation phase of the system is extracted, and a baseline health point is established in the multi-dimensional state-space model. By calculating the spatial distance between the real-time state point and the baseline health point in the multi-dimensional state-space model, the severity of the current state deviating from the health baseline is quantified, and then the comprehensive health index of each carbon brush is determined based on the magnitude of the spatial distance.

[0024] More specifically, in the stage of generating early warning commands, the state fusion and early warning module is equipped with a logical evaluation step to distinguish different types of deteriorating working conditions. When the electrical contact state parameters are detected to increase gradually and slowly, while the pure mechanical frictional heat parameters remain within the normal range, it is determined that the carbon brush is in a normal decay state of spring pressure during long-term operation. When the electrical contact state parameters are detected to change abruptly and abruptly, accompanied by pure mechanical frictional heat parameters that significantly exceed the normal range, it is determined that the carbon brush is in a state of mechanical jamming in the brush holder or sudden spring breakage. When the micro-arc discharge parameters are detected to be continuously non-zero and show a significant upward trend, it is determined that the interface between the carbon brush and the slip ring can no longer maintain stable electrical contact and has entered a state of severe electrical ablation.

[0025] Furthermore, this embodiment also includes a lifespan prediction module for proactive management of the entire equipment lifecycle. The lifespan prediction module acquires historical cumulative values ​​of pure mechanical frictional heat parameters and maps them equivalently to mechanical wear reflecting physical consumption; simultaneously, it acquires historical cumulative values ​​of micro-arc discharge parameters and maps them equivalently to electrical ablation wear reflecting electrical spark erosion. The mechanical wear and electrical ablation wear are then substituted into a preset multi-factor wear accumulation model. Based on the current degradation slope, the remaining time required for the comprehensive health index to reach the preset scrap threshold is predicted, and the remaining service life of each carbon brush is dynamically output, providing data support for on-site maintenance personnel to formulate scientific maintenance plans.

[0026] Furthermore, to prevent false alarms from being triggered by the generator under extreme load conditions, this system introduces a dynamic adaptive threshold mechanism. The data acquisition module also simultaneously acquires the generator's global load data. The status fusion and early warning module, based on the generator's global load data from historical normal operation phases and the corresponding comprehensive health indicators, trains a dynamic threshold model using machine learning algorithms. In the real-time monitoring phase, the current generator global load data is input into the dynamic threshold model, which dynamically outputs an allowable health safety boundary that matches the current load condition. An early warning command is only triggered when the real-time comprehensive health indicator exceeds the health safety boundary. By dynamically adjusting the early warning limits, the system significantly reduces the false alarm rate while maintaining high sensitivity.

[0027] The system's warning commands are categorized into three levels based on the urgency and severity of equipment degradation. When the overall health index is in the first degradation range, it indicates a minor system anomaly or normal wear approaching the limit. At this point, a Level 1 warning is issued, suggesting scheduled cleaning or regular maintenance. When the overall health index further deteriorates to the second degradation range, and the pure mechanical frictional heat parameters exceed the limit, it indicates severe single-point physical heating. At this point, a Level 2 warning is issued, suggesting adjusting the generator's reactive power output or reducing the cooling air temperature to alleviate thermal stress on the equipment. When the overall health index falls into the third degradation range, and the micro-arc discharge parameters continuously exceed the limit, potentially triggering a catastrophic arcing accident, a Level 3 warning is issued, directly triggering the generator's emergency trip protection or cutting off the excitation circuit to maximize the overall safety of the generator set.

[0028] Example 2 Based on the same inventive concept as the aforementioned online monitoring and early warning system for generator carbon brush condition, another embodiment of the present invention provides an online monitoring and early warning method for generator carbon brush condition. This embodiment can be applied to electronic devices with communication, computing, and data storage capabilities. The specific process is as follows: Figure 2 As shown, it includes: The operating status data of each carbon brush on the same polarity brush holder of the generator is obtained. The operating status data includes shunt current data, temperature data and high frequency characteristic data.

[0029] Specifically, operational status data refers to multi-dimensional physical quantities reflecting the real-time electromagnetic and thermodynamic conditions of carbon brushes. Operational status data can be accurately captured using distributed sensing and multi-channel synchronous acquisition technology. Multiple carbon brushes are typically connected in parallel on the same polarity brush holder of a generator. Different sampling logics are employed at the data acquisition end to address the frequency band characteristics of different physical signals. For shunt current and temperature data, a low-frequency sampling channel is used for timed acquisition to capture quasi-static trends that change slowly with generator load and environment. For high-frequency characteristic data, including high-frequency current and wideband vibration, a high-frequency sampling channel is used for event-triggered acquisition to capture transient abnormal signals within a very short time window. Furthermore, the data from the low-frequency and high-frequency sampling channels are timestamped and aligned using the system bus clock to generate operational status data with a unified time reference.

[0030] An equivalent circuit model of the carbon brush is constructed based on the shunt current data to obtain the electrical contact state parameters of each carbon brush.

[0031] Specifically, since directly measuring the resistance of the carbon brush-slip ring contact surface is extremely difficult, this step aims to invert the contact state through circuit topology reconstruction. The generator slip ring body is used as the common reference terminal for current collection, and the busbar of the same polarity brush holder is used as the common injection terminal. Each parallel carbon brush and its interface with the slip ring constitutes multiple parallel time-varying impedance branches, thus constructing an equivalent circuit model of the carbon brush. The acquired shunt current data is substituted into the equivalent circuit model of the carbon brush for impedance calculation. Based on the dynamic distribution ratio of the current in each branch, the real-time equivalent contact resistance of each branch is analyzed in reverse, and this real-time equivalent contact resistance is used as an electrical contact state parameter. The variation law of the real-time equivalent contact resistance objectively reflects the electrical conduction quality of the contact interface between the bottom surface of the carbon brush and the slip ring.

[0032] Based on temperature and shunt current data, a thermal-electric decoupling model is constructed to separate the pure mechanical frictional heat parameters of each carbon brush.

[0033] Specifically, in actual generator operation, the temperature rise of the carbon brushes is the result of the combined effects of Joule heating from the current and mechanical frictional heat. A thermo-electric decoupling model is used to isolate electromagnetic heating interference from the mixed thermal field. An electrothermal calculation model is constructed based on shunt current data, and the electrothermal component caused solely by the current passing through each carbon brush is calculated according to the Joule heating effect. The background ambient temperature inside the generator is extracted. By subtracting the corresponding electrothermal component and the background ambient temperature from the real-time temperature data of each carbon brush, the heat generated purely by mechanical friction during high-speed relative sliding between the carbon brush and the slip ring is separated and used as the pure mechanical frictional heat parameter. The extraction of the pure mechanical frictional heat parameter provides a clean physical quantity basis for accurately assessing the mechanical jamming state and spring clamping force state inside the brush holder.

[0034] Based on high-frequency characteristic data, a time-frequency joint analysis model was constructed to extract the micro-arc discharge parameters of each carbon brush.

[0035] Specifically, when a micro-arc discharge occurs, it not only induces high-frequency oscillating current in the electrical circuit but also triggers a microscopic acoustic explosion impact at the contact interface. The time-frequency joint analysis model aims to eliminate electromagnetic noise interference from the external space of the generator. The energy envelopes of the high-frequency current signal and the broadband vibration signal within a set high-frequency band are extracted separately. The cross-correlation coefficient between the energy envelopes of the high-frequency current signal and the broadband vibration signal is calculated. A genuine micro-arc discharge phenomenon is confirmed only if the cross-correlation coefficient is greater than a preset threshold. After confirming a genuine discharge, the energy envelope of the high-frequency current signal is integrated to evaluate the total energy released in a single or continuous discharge, and this is used as a parameter for the micro-arc discharge.

[0036] The electrical contact state parameters, pure mechanical frictional heat parameters, and micro-arc discharge parameters are fused into a model to obtain a comprehensive health index for each carbon brush, and an early warning command is output based on the comprehensive health index.

[0037] Specifically, a multidimensional state-space model is constructed, with electrical contact state parameters, pure mechanical frictional heat parameters, and micro-arc discharge parameters as independent dimensions. Real-time parameters of each carbon brush are mapped into the multidimensional state-space model, forming real-time state points reflecting the current operating state of the carbon brushes. Simultaneously, historical baseline data from the initial healthy operation phase of the system is extracted, and a baseline health point is established in the multidimensional state-space model. By calculating the spatial distance between the real-time state point and the baseline health point in the multidimensional state-space model, the severity of the current state deviating from the health baseline is quantified, thereby determining the comprehensive health index of each carbon brush. In the early warning command generation stage, when the electrical contact status parameters are detected to increase gradually and slowly, while the pure mechanical frictional heat parameters remain within the normal range, the carbon brush is determined to be in a normal decay state of spring pressure during long-term operation. When the electrical contact status parameters are detected to change abruptly and abruptly, accompanied by pure mechanical frictional heat parameters significantly exceeding the normal range, the carbon brush is determined to be in a state of mechanical jamming within the brush holder or sudden spring breakage. When the micro-arc discharge parameters are detected to be continuously non-zero and showing a significant upward trend, the carbon brush and slip ring interface is determined to be unable to maintain stable electrical contact and has entered a state of severe electrical ablation. Based on the different determined states, a graded early warning command is output, including prompts for scheduled cleaning, suggestions for adjusting generator reactive power output, or linkage of generator emergency trip protection.

[0038] In summary, the online monitoring and early warning method for generator carbon brush status provided in this embodiment can meticulously depict the dynamic changes in the conductivity of the carbon brush body by acquiring multi-dimensional operating status data and constructing an equivalent circuit model of the carbon brush based on shunt current data. By constructing a thermal-electric decoupling model based on temperature and current data, electromagnetic heating interference is accurately isolated, enabling the sensitive detection of thermal distortion caused by purely mechanical faults. Micro-arc discharge parameters are extracted based on high-frequency feature data, overcoming the limitation of traditional optical equipment in capturing hidden sparks. The fusion of multi-physics parameters into a model not only enables early warning of abnormal states but also accurately locates the cause of faults, greatly improving the monitoring's anti-interference capability and prediction accuracy. The final output warning command can maximize the overall safety of the generator set.

[0039] Example 3 Based on the same inventive concept as the aforementioned online monitoring and early warning system for generator carbon brush status, another embodiment of the present invention provides an electronic device. This electronic device is intended to serve as the physical computing carrier for the aforementioned monitoring and early warning architecture, achieving the processing of multi-dimensional operational data and the decoupling computation of multi-physics models through underlying hardware resource scheduling.

[0040] Specifically, the electronic equipment mainly includes a processor, memory, and communication interface connected via a system bus. The communication interface is used to exchange data with various high- and low-frequency sensors on the local side of the generator and the distributed control system of the remote control center to ensure real-time access to the generator's global operating condition data and the micro-operating status data of the carbon brush.

[0041] As the central storage hub for non-volatile or volatile data, memory stores computer programs and various historical baseline data generated during model computation. Memory can take the form of read-only memory, random access memory, programmable read-only memory, or flash memory, providing sufficient operating memory and data persistence space for complex multidimensional state-space models and particle filter lifetime prediction models.

[0042] The processor, as the core control and computing engine of the electronic device, is used to read and execute computer programs stored in memory. When executing the computer program, the processor specifically implements the following in-depth logical deduction control: It acquires the operating status data of each carbon brush on the same polarity brush holder of the generator, including shunt current data, temperature data, and high-frequency characteristic data, via the communication interface; it schedules computing resources to abstract the slip ring and carbon brush array into an equivalent circuit model, and solves the real-time equivalent contact resistance of the bottom surface of each carbon brush through Kirchhoff matrix iteration; it introduces the Joule law electrothermal calculation system and combines it with the generator background ambient temperature to extract the pure mechanical frictional heat parameters that accurately reflect the brush box jamming and spring state from the macroscopic mixed temperature data; it calls high-frequency signal processing instructions to extract the envelope of high-frequency current signals and broadband vibration signals and perform time-domain cross-correlation verification to avoid spatial electromagnetic noise interference and quantify the real micro-arc discharge energy; finally, it constructs a multi-dimensional state space coordinate system with electrical contact state parameters, pure mechanical frictional heat parameters, and micro-arc discharge parameters as dimensions, generates a comprehensive health index by calculating the spatial distance of the real-time state point from the benchmark health point, and adaptively outputs graded early warning instructions for maintenance scheduling, flexible operating condition adjustment, or hard interlock tripping based on a pre-trained dynamic threshold model.

[0043] Understandably, the processor adopts a chip architecture with high-performance floating-point operation and matrix solving capabilities, such as a central processing unit, microprocessor, digital signal processor or application-specific integrated circuit, to ensure that time-frequency joint analysis and discharge feature extraction can meet the system's stringent requirements for extremely low latency in micro-arc early warning when high-frequency transient signals arrive.

[0044] Example 4 Based on the same inventive concept as the aforementioned system embodiments, another embodiment of the present invention provides a computer-readable storage medium. This computer-readable storage medium aims to explicitly solidify the implicit algorithmic logic of the aforementioned system's multidimensional physical field decoupling and adaptive early warning at the code level.

[0045] Specifically, a computer-readable storage medium stores a computer program. When the computer program is executed by a computer device or a microcontroller unit with data processing capabilities, it can accurately reproduce all the logical steps defined by the aforementioned online monitoring and early warning system for generator carbon brush status. That is, by executing the computer program, the computer device can synchronously capture carbon brush operating data across physical fields and strictly follow the data flow to sequentially drive the thermoelectric decoupling model to separate mechanical frictional heat, drive the time-frequency joint analysis model to confirm hidden micro-arcs, and drive the multi-dimensional state-space model to assess overall health. The implementation of this storage medium breaks through the technical bottleneck of traditional monitoring systems that rely excessively on physical hardware, and performs high-precision mathematical mapping and closed-loop early warning of the evolution law of carbon brush groups under multi-dimensional complex operating conditions through a pure software algorithm architecture. In some optional embodiments, the computer-readable storage medium can be a USB flash drive, external hard drive, magnetic disk, optical disk, or a distributed data block deployed on a cloud server, and is not limited to a specific physical carrier form.

[0046] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. An online monitoring and early warning system for the condition of a generator carbon brush, characterized in that, include: The data acquisition module is used to acquire the operating status data of each carbon brush on the same polarity brush holder of the generator. The operating status data includes shunt current data, temperature data, and high-frequency characteristic data. The contact state modeling module, connected to the data acquisition module, is used to construct an equivalent circuit model of the carbon brush based on the shunt current data, and obtain the electrical contact state parameters of each carbon brush. The heat source separation modeling module is connected to the data acquisition module and is used to construct a thermal-electric decoupling model based on the temperature data and shunt current data to separate and obtain the pure mechanical frictional heat parameters of each carbon brush. The discharge characteristic modeling module is connected to the data acquisition module and is used to construct a time-frequency joint analysis model based on the high-frequency characteristic data to extract the micro-arc discharge parameters of each carbon brush. The state fusion and early warning module is connected to the contact state modeling module, the heat source separation modeling module, and the discharge characteristic modeling module, respectively. It is used to fuse the electrical contact state parameters, pure mechanical friction heat parameters, and micro-arc discharge parameters into a model to obtain the comprehensive health index of each carbon brush, and output an early warning command based on the comprehensive health index.

2. The generator carbon brush condition online monitoring and early warning system according to claim 1, characterized in that, The contact state modeling module performs the following steps: The generator slip ring and the parallel carbon brushes on the same polarity brush holder are used to construct the equivalent circuit model of the carbon brush. The shunt current data of each carbon brush is used as an input variable and substituted into the equivalent circuit model of the carbon brush to solve for impedance, thereby obtaining the real-time equivalent contact resistance of each carbon brush, and the real-time equivalent contact resistance is used as the electrical contact state parameter.

3. The generator carbon brush condition online monitoring and early warning system according to claim 1, characterized in that, The specific execution steps of the heat source separation modeling module are as follows: An electrothermal calculation model is constructed based on the shunt current data of each carbon brush to calculate the electrothermal component caused by the current in each carbon brush. Extract the background ambient temperature inside the generator; Subtract the corresponding electrothermal component and background ambient temperature from the temperature data of each carbon brush to separate the heat generated by mechanical friction between the carbon brush and the slip ring, and use it as the pure mechanical frictional heat parameter.

4. The generator carbon brush condition online monitoring and early warning system according to claim 1, characterized in that, The discharge characteristic modeling module performs the following steps: The high-frequency characteristic data includes high-frequency current signals and wideband vibration signals; The energy envelopes of the high-frequency current signal and the wideband vibration signal within a set high-frequency band are extracted respectively. Calculate the cross-correlation coefficient between the energy envelope of a high-frequency current signal and the energy envelope of a broadband vibration signal; When the cross-correlation coefficient is greater than a preset threshold, a real micro-arc discharge is confirmed, and the energy integral value of the high-frequency current signal is calculated and used as the micro-arc discharge parameter.

5. The generator carbon brush condition online monitoring and early warning system according to claim 1, characterized in that, The specific execution steps of the state fusion and early warning module are as follows: A multidimensional state-space model is constructed with electrical contact state parameters, pure mechanical frictional heat parameters, and micro-arc discharge parameters as dimensions. The real-time parameters of each carbon brush are mapped to a multi-dimensional state space model to form real-time state points; Calculate the spatial distance between the real-time status point and the baseline health point in the multidimensional state space model, and determine the comprehensive health index of each carbon brush based on the magnitude of the spatial distance.

6. The generator carbon brush condition online monitoring and early warning system according to claim 1 or 5, characterized in that, The state fusion and early warning module further includes the following specific execution steps: When the electrical contact state parameters are detected to be gradually increasing and the pure mechanical friction heat parameters are within the set normal range, the carbon brush is determined to be in a normal state of spring pressure decay. When a step change is detected in the electrical contact status parameters and the pure mechanical frictional heat parameters exceed the set normal range, the carbon brush is determined to be stuck or the spring is broken. When the micro-arc discharge parameters are continuously non-zero and show an upward trend, the carbon brush is determined to be in an electro-ablation state.

7. The generator carbon brush condition online monitoring and early warning system according to claim 1, characterized in that, It also includes a lifetime prediction module, which is used for: Obtain the historical cumulative values ​​of pure mechanical frictional heat parameters and convert them into mechanical wear. The historical cumulative values ​​of micro-arc discharge parameters are obtained and converted into electrical ablation wear. By substituting the mechanical wear and electrical ablation wear into a preset wear accumulation model, the remaining time before the comprehensive health index reaches the scrap threshold is predicted, and the remaining service life of the carbon brush is output.

8. A method for online monitoring and early warning of generator carbon brush condition, characterized in that, include: The operating status data of each carbon brush on the same polarity brush holder of the generator is obtained, and the operating status data includes shunt current data, temperature data and high frequency characteristic data. Based on the shunt current data, an equivalent circuit model of the carbon brush is constructed to obtain the electrical contact state parameters of each carbon brush. Based on the temperature data and shunt current data, a thermal-electric decoupling model is constructed to separate the pure mechanical frictional heat parameters of each carbon brush. Based on the high-frequency characteristic data, a time-frequency joint analysis model is constructed to extract the micro-arc discharge parameters of each carbon brush. The electrical contact state parameters, pure mechanical frictional heat parameters, and micro-arc discharge parameters are fused into a model to obtain a comprehensive health index for each carbon brush, and an early warning command is output based on the comprehensive health index.

9. An electronic device, characterized in that, include: At least one processor; as well as A memory that is communicatively connected to the at least one processor; The memory stores a computer program that can be executed by the at least one processor, which enables the at least one processor to perform the functions of the generator carbon brush condition online monitoring and early warning system as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions for causing the computer to perform the functions of the generator carbon brush condition online monitoring and early warning system as described in any one of claims 1 to 7.

Images