Oil-to-electricity carbon brush height identification early warning system

Abstract

The invention relates to the field of gantry crane sliding contact line carbon brush monitoring and intelligent early warning, in particular to an oil-to-electricity carbon brush height identification early warning system, which adopts various sensors to acquire temperature, vibration, images and other data of a carbon brush and a sliding contact line in real time. A temperature field, a stress field and wear rate data are generated through a data preprocessing and multi-dimensional thermal-mechanical coupling modeling method; the system is combined with an incremental learning module to continuously optimize the prediction capability of the model; through state evaluation of the carbon brush and the slide wire, the system can generate risk scoring data in real time, perform fault prediction, and trigger automatic shutdown protection or load adjustment measures under a high-risk condition. Through accurate fault identification and risk early warning, the problems of data lag and missing detection in a traditional monitoring method are effectively reduced, and the safety, reliability and intelligent management level of equipment are improved.

Description

Technical Field

[0001] The invention relates to the field of gantry crane busbar carbon brush monitoring and intelligent early warning, and in particular to an oil-to-electric carbon brush height identification and early warning system. Background Art

[0002] With the development of automation and intelligence in modern logistics and port operations, the status monitoring of carbon brushes on gantry crane busbars has become increasingly important. Gantry cranes are important mechanical equipment in ports, cargo yards and other places, and are widely used for lifting and stacking of goods. The busbar is a key component of the gantry crane's power supply system. The contact status between the carbon brush and the busbar is directly related to the stable transmission of power and the normal operation of the equipment. Problems such as wear, overheating, and vibration of the carbon brush often lead to equipment failures and even shutdowns, which have a serious impact on production. Therefore, real-time monitoring and fault prediction of carbon brushes and busbars are particularly critical.

[0003] At present, most traditional carbon brush monitoring methods rely on manual inspections or simple sensor systems, which determine whether there is a problem by checking single data such as the appearance and temperature of the carbon brush. However, these traditional methods have many shortcomings. First, manual inspections are inefficient. Limited by the experience and work frequency of inspectors, details are easily overlooked, and the judgment of equipment status lacks timeliness and accuracy. Secondly, existing sensors mostly rely on a single data source, such as temperature sensors or vibration sensors, and cannot comprehensively consider the multi-dimensional status of carbon brushes (such as height, temperature, vibration, etc.), resulting in limitations and incompleteness of monitoring results. In addition, traditional methods fail to predict the wear rate and potential failures of carbon brushes, and often only take measures after the equipment fails, causing additional damage to the equipment and maintenance costs. Summary of the invention

[0004] In view of the many problems existing in the above-mentioned prior art, the present invention provides an oil-to-electric carbon brush height identification and warning system. The present invention integrates a variety of sensors (such as temperature, vibration, image, etc.) to collect the operating data of carbon brushes and busbars in real time, and adopts a multi-dimensional thermal-mechanical coupling modeling method for data analysis and wear prediction. At the same time, through the incremental learning module, the system can continuously optimize the prediction model, and the prediction accuracy continues to improve over time. Through accurate fault prediction and real-time monitoring, the present invention can identify equipment risks in advance, trigger warnings and execute protective measures (such as automatic shutdown or load adjustment), effectively reducing the probability of failure and improving the safety and reliability of equipment operation.

[0005] A oil-to-electric carbon brush height recognition and warning system, comprising: A data acquisition unit, used to collect carbon brush height data, temperature data, vibration data, busbar surface image data and environmental parameter data; A data preprocessing unit, used for preprocessing the collected data to generate fused data; A modeling unit, for calculating and generating temperature field data, stress field data and wear rate prediction data based on the fused data by adopting a multi-dimensional thermo-mechanical coupling modeling method, and extracting joint feature data; The analysis unit is used to perform nonlinear analysis on the joint feature data to generate risk scoring data, and perform risk assessment on the operating status of the carbon brush and the busbar according to the risk scoring data, thereby triggering an early warning or executing protective measures.

[0006] Preferably, the data acquisition unit comprises: Laser distance sensor, used to collect the height data of carbon brush; Infrared thermal imager, used to collect temperature data of carbon brushes; High-frequency vibration sensor, used to collect vibration data of carbon brushes; Visual depth camera, used to collect busbar surface image data; Humidity and air pressure sensors are used to collect environmental parameter data.

[0007] Preferably, the data preprocessing unit includes: A data formatting module, used to convert the carbon brush height data, carbon brush temperature data, carbon brush vibration data, busbar surface image data and environmental parameter data collected from different sensors into a unified format; Noise filtering module, used to remove noise from the collected data to ensure the accuracy of the data; The time alignment module is used to time align the carbon brush height data, carbon brush temperature data, carbon brush vibration data, busbar surface image data and environmental parameter data collected by different sensors to generate fused data.

[0008] Preferably, the modeling unit comprises: A multi-dimensional thermal-mechanical coupling module, which is used to mesh the contact area between the carbon brush and the busbar using the finite element method based on the fusion data, and calculate and generate temperature field data and stress field data according to the interaction between the temperature field and the stress field in the contact area between the carbon brush and the busbar. The temperature field data is used to describe the heat distribution generated during the friction process, and the stress field data is used to describe the contact force between the carbon brush and the busbar. The stress field data is further used to evaluate the pressure distribution of the contact point; A wear rate prediction module is used to calculate the wear rate of the carbon brush according to the temperature field data and the stress field data. The wear rate prediction module is based on Archard's law and predicts the wear rate of the carbon brush during use by analyzing the contact force, temperature distribution and contact duration between the carbon brush and the busbar, and generates the remaining service life of the carbon brush; The feature extraction module is used to extract joint feature data from the temperature field data, stress field data and wear rate prediction data. The feature extraction module uses a principal component analysis algorithm to reduce the dimension of multidimensional data and compress redundant feature information into a low-dimensional feature vector.

[0009] Preferably, the analysis unit comprises: A nonlinear analysis module, used for performing nonlinear analysis on the joint feature data, wherein the nonlinear analysis module is based on a multi-layer perceptron neural network, uses a supervised learning method to train the joint feature data, and generates risk score data through nonlinear mapping, wherein the risk score data represents the potential failure risk of the carbon brush and busbar system, and the neural network includes an input layer, multiple hidden layers and an output layer, and uses a back propagation algorithm to update weights and minimize errors; The risk assessment module is used to perform risk assessment on the operating status of the carbon brush and the busbar according to the risk scoring data. The risk assessment module determines the risk level of the carbon brush and the busbar according to the set threshold value. The risk level includes low risk, medium risk and high risk. The risk assessment result determines whether to trigger an early warning or start protective measures.

[0010] Preferably, when the risk level is high, the risk assessment module triggers protective measures, including: Automatic shutdown protection, used to stop the operation of carbon brushes or busbar equipment when high risk is detected. The automatic shutdown protection is connected to the equipment power supply through the control system to interrupt the power supply of the equipment in real time to prevent further damage to the equipment; Adjusting the load distribution is used to reduce the risk of failure by controlling the load distribution of the busbar when the risk assessment is medium risk. The load distribution is adjusted by adjusting the load balancing module to ensure reasonable distribution of the busbar load and reduce equipment wear.

[0011] Preferably, the system further comprises: A storage unit, used to store the carbon brush height data, carbon brush temperature data, carbon brush vibration data, busbar surface image data and environmental parameter data, the storage unit adopts a high-performance hard disk or solid-state hard disk and supports real-time data backup; The incremental learning module is used to optimize the computing power of the modeling unit and the analysis unit based on historical data and real-time data. The incremental learning module updates the model through regular training, so that the system can continuously improve the prediction accuracy according to new data.

[0012] Preferably, the data acquisition unit is connected to the data preprocessing unit, the modeling unit and the analysis unit via a communication network. The communication network uses an industrial-grade Ethernet protocol, is capable of transmitting the collected real-time data to each unit, and returns the processing results to the user or the control system. The transmission delay of the communication network does not exceed 10 milliseconds.

[0013] Preferably, the system further comprises: The display unit is used to display the current status, risk level and processing suggestions of the carbon brush and the busbar. The display unit adopts a touch screen interface and supports a real-time refresh function for reference and operation by maintenance personnel.

[0014] Compared with the prior art, the advantages and beneficial effects of the present invention are: Through multi-dimensional data fusion and intelligent modeling and analysis technology, the present invention realizes real-time collection, analysis and early warning of multi-dimensional data (such as carbon brush height, temperature, vibration, etc.) of carbon brushes and busbars; Compared with the existing traditional technology, the present invention not only improves the accuracy and comprehensiveness of data collection, but also can predict faults through intelligent algorithms combined with real-time data, and warn of equipment failures in advance, thereby avoiding production stagnation and equipment damage caused by carbon brush wear or failure; Through incremental learning and nonlinear analysis methods, the present invention can continuously optimize the model so that the system can continuously improve the accuracy of fault prediction based on new data, thereby further reducing the failure rate and maintenance costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 It is a structural block diagram of the system of the present invention; Figure 2 It is a flow chart of data flow and relationship between modules in the present invention; Figure 3 It is a schematic diagram of the modeling and analysis module in the present invention; Figure 4 It is a schematic diagram of the risk assessment and decision triggering module in the present invention; Figure 5 This is a flow chart for triggering risk assessment and protective measures in the present invention. DETAILED DESCRIPTION

[0016] Hereinafter, embodiments of the present disclosure will be described with reference to the accompanying drawings. However, it should be understood that these descriptions are exemplary only and are not intended to limit the scope of the present disclosure. In the following detailed description, for ease of explanation, many specific details are set forth to provide a comprehensive understanding of the embodiments of the present disclosure. However, it is obvious that one or more embodiments may also be implemented without these specific details. In addition, in the following description, descriptions of known structures and technologies are omitted to avoid unnecessary confusion of the concepts of the present disclosure.

[0017] The terms used herein are only for describing specific embodiments and are not intended to limit the present disclosure. The terms "comprise", "include", etc. used herein indicate the existence of the features, steps, operations and / or components, but do not exclude the existence or addition of one or more other features, steps, operations or components.

[0018] All terms (including technical and scientific terms) used herein have the meanings commonly understood by those skilled in the art unless otherwise defined. It should be noted that the terms used herein should be interpreted as having a meaning consistent with the context of this specification and should not be interpreted in an idealized or overly rigid manner.

[0019] like Figure 1-Figure 2 As shown, a oil-to-electric carbon brush height recognition and warning system comprises: A data acquisition unit, used to collect carbon brush height data, temperature data, vibration data, busbar surface image data and environmental parameter data; Preferably, the data acquisition unit comprises: Laser distance sensor, used to collect the height data of carbon brush; Infrared thermal imager, used to collect temperature data of carbon brushes; High-frequency vibration sensor, used to collect vibration data of carbon brushes; Visual depth camera, used to collect busbar surface image data; Humidity and air pressure sensors are used to collect environmental parameter data.

[0020] The laser distance sensor measures the precise height of the carbon brush relative to the ground by emitting a laser beam and receiving the reflected signal. Its working principle is based on the time difference of the laser pulse reflection (time of flight method), which can accurately obtain the height change of the carbon brush during operation.

[0021] In the present invention, the laser distance sensor is used to collect the height data of the carbon brush in real time, in millimeters, with an accuracy of ±0.1 mm. The sampling frequency of the sensor is 10 Hz, that is, data is collected 10 times per second, ensuring that the height data can be obtained at a high frequency during the contact between the carbon brush and the busbar. By continuously monitoring the changes in the height of the carbon brush, the wear, deformation or offset of the carbon brush can be discovered in time, thereby triggering the corresponding early warning and protection mechanism.

[0022] The infrared thermal imager generates a temperature image by capturing infrared radiation in the contact area between the carbon brush and the busbar. Its working principle is to infer the temperature distribution through the change in the amount of thermal radiation. The infrared thermal imager has a strong adaptability to inaccessible areas (such as the contact surface between the inside of the carbon brush and the busbar), and can monitor temperature changes in real time and generate a two-dimensional temperature distribution matrix.

[0023] In the present invention, the infrared thermal imager is used to collect the temperature data of the carbon brush. Specifically, the collected temperature data is a two-dimensional temperature distribution matrix, in degrees Celsius, with a resolution of 0.1°C, which can accurately describe the temperature change in the contact area between the carbon brush and the busbar. By analyzing the temperature distribution of the carbon brush in real time, the system can detect overheating areas, thereby preventing equipment damage or failure caused by excessive temperature.

[0024] The high-frequency vibration sensor obtains the vibration data of the carbon brush by detecting the vibration signal generated in the contact area between the carbon brush and the busbar. These vibration signals are usually caused by friction, impact or irregular contact between the carbon brush and the busbar. The sensor converts the vibration signal into an electrical signal and compares it with the time series data to analyze the working status of the carbon brush.

[0025] In the present invention, vibration data is collected in the form of acceleration time series signals, with the unit of g (gravitational acceleration), the sampling frequency of 1kHz, and the accuracy of ±0.01g. Through high-frequency sampling, the subtle vibration changes of the carbon brush can be accurately captured, and abnormal vibrations, such as excessive vibrations caused by carbon brush damage or instability, can be discovered in time. These vibration data can be used for further fault diagnosis to help the system detect early wear or other potential problems.

[0026] The visual depth camera combines traditional RGB images with depth map imaging technology to simultaneously obtain the appearance information and depth information of the contact area between the carbon brush and the busbar. Its working principle is to project a specific pattern on the surface of an object through laser or structured light technology. The camera captures the reflected image and calculates the depth information of the object surface, thereby constructing the three-dimensional structure of the area.

[0027] In the present invention, the visual depth camera is used to collect image data of the busbar surface, providing RGB images and depth maps, with an image resolution of 1920×1080, a depth map resolution of 1280×720, and a depth accuracy of ±1 mm. Through the image data, the system can monitor the contact surface between the carbon brush and the busbar in real time, and detect damage, wear or other abnormal conditions on the busbar surface. Combined with temperature data and vibration data, the image data provides a more intuitive basis for the status analysis of the carbon brush.

[0028] Humidity and air pressure sensors are used to monitor changes in humidity and air pressure in the environment. These two environmental parameters have an important impact on the performance of carbon brushes and busbars. Excessive humidity may cause corrosion of electrical components, while changes in air pressure may affect the operating efficiency and stability of the equipment.

[0029] In the present invention, humidity and air pressure sensors are used to collect environmental data in real time. The humidity data unit is percentage, and the relative humidity accuracy is ±1%. The air pressure data unit is hPa, and the air pressure accuracy is ±0.5hPa. These data can help the system dynamically adjust the working parameters according to environmental conditions to ensure the normal operation of the carbon brush and busbar system in different environments.

[0030] A data preprocessing unit, used for preprocessing the collected data to generate fused data; Preferably, the data preprocessing unit includes: A data formatting module, used to convert the carbon brush height data, carbon brush temperature data, carbon brush vibration data, busbar surface image data and environmental parameter data collected from different sensors into a unified format; Noise filtering module, used to remove noise from the collected data to ensure the accuracy of the data; The time alignment module is used to time align the carbon brush height data, carbon brush temperature data, carbon brush vibration data, busbar surface image data and environmental parameter data collected by different sensors to generate fused data.

[0031] The main function of the data formatting module is to convert the raw data collected from different sensors into a standard format for subsequent processing and analysis. Since the data collected by different types of sensors (such as laser ranging sensors, infrared thermal imagers, vibration sensors, etc.) often differ in terms of units, data structures, time stamps, etc., formatting must be performed to ensure data compatibility and consistency.

[0032] Specifically, the carbon brush height data is usually in millimeters, and it is necessary to ensure that all height data are converted with the same accuracy (±0.1mm); the temperature data is in Celsius, and the temperature data is required to be processed with standard accuracy (0.1℃); the vibration data is expressed in acceleration units (g), and the sampling frequency and unit need to be unified, usually collected at a frequency of 1kHz. Image data will be standardized in size and color space so that it can be synchronized with other data; environmental data (humidity and air pressure) need to be unified in units and processed into an analyzable format.

[0033] This module implements the above formatting by defining a standardized data format (such as JSON, CSV, or a standard database table). This process involves not only unit conversion, but also the unification of data precision so that subsequent modules can process this data.

[0034] The data formatting module ensures that data collected by different types of sensors are compatible and effectively processed. This preprocessing step allows all data types (temperature, vibration, image, carbon brush height, etc.) to be stored in a unified format, providing a reliable data foundation for subsequent analysis and modeling. The formatted data provides accurate input for the model, thereby improving the accuracy and efficiency of data analysis.

[0035] The core function of the noise filtering module is to remove noise from the data and ensure that the collected signals are as close to the real physical phenomenon as possible. Due to environmental factors, equipment accuracy limitations, sensor errors and other reasons, the raw data collected by the sensor often contains noise. If these noises are not processed, they may affect subsequent data analysis and lead to inaccurate model training and risk assessment.

[0036] The noise filtering module usually adopts an adaptive Kalman filter algorithm, which is widely used in dynamic signal processing. The Kalman filter automatically estimates the real signal and suppresses random noise by combining the predicted value of the current signal with the real-time collected data. In the present invention, the core of the Kalman filter algorithm is to adjust the filter parameters according to the dynamic changes of the measurement error, so as to filter out the noise more accurately and retain important signal features.

[0037] For example, in the carbon brush vibration data processing, the Kalman filter automatically adjusts the weight according to the measurement error of each sampling cycle, thereby reducing the impact of sensor error and environmental interference, ensuring that the vibration signal reflects the true working state of the carbon brush. Similarly, the carbon brush temperature data and image data will also use similar filtering technology to remove environmental interference and sensor errors.

[0038] Through noise filtering, the collected data such as carbon brush vibration, temperature, image, etc. will be more accurate, noise interference is effectively suppressed, and the signal-to-noise ratio of the data is significantly improved. This provides a more reliable data basis for subsequent modeling and nonlinear analysis, ensuring that the system's monitoring of the status of carbon brushes and busbars can reflect the actual situation and avoid misjudgment or missed potential faults caused by noise.

[0039] The main task of the time alignment module is to synchronize the data from different sensors according to the timestamp. Since the working principles and sampling frequencies of each sensor are different, the data they collect usually have different time bases, so the data needs to be aligned to the same time axis.

[0040] In the present invention, time alignment is accomplished through precise clock synchronization technology. The sampled data of each sensor will be accompanied by a timestamp, which indicates the actual time of data acquisition. During data preprocessing, the time alignment module aligns the time of different data sources (such as carbon brush height, temperature, vibration, image, etc.) to a unified time base based on the timestamp of the sensor data acquisition, ensuring that all data points are consistent with their corresponding timestamps. Specifically, the interpolation method is used in the time alignment process, and the sampling points are interpolated so that data of different time scales can correspond accurately, thereby achieving time synchronization of the data.

[0041] For example, the carbon brush vibration data and the carbon brush height data may have slight differences in time. The time alignment module will automatically adjust their time axes so that these data can be compared at the same time point, thereby ensuring that the model uses synchronized time series data during training.

[0042] The function of the time alignment module is to eliminate the time differences of different sensor data so that all data can be accurately arranged in chronological order. This process provides a reliable foundation for subsequent data fusion, ensuring that the model can simultaneously consider the information of various data sources at the same time point during analysis, thereby improving the accuracy and responsiveness of the system. Especially in modeling and risk assessment, time-aligned data allows the system to make a more accurate analysis of the interaction between carbon brushes and busbars.

[0043] A modeling unit, for calculating and generating temperature field data, stress field data and wear rate prediction data based on the fused data by adopting a multi-dimensional thermo-mechanical coupling modeling method, and extracting joint feature data; Preferably, Figure 3 As shown, the modeling unit includes: A multi-dimensional thermal-mechanical coupling module, which is used to mesh the contact area between the carbon brush and the busbar using the finite element method based on the fusion data, and calculate and generate temperature field data and stress field data according to the interaction between the temperature field and the stress field in the contact area between the carbon brush and the busbar. The temperature field data is used to describe the heat distribution generated during the friction process, and the stress field data is used to describe the contact force between the carbon brush and the busbar. The stress field data is further used to evaluate the pressure distribution of the contact point; A wear rate prediction module is used to calculate the wear rate of the carbon brush according to the temperature field data and the stress field data. The wear rate prediction module is based on Archard's law and predicts the wear rate of the carbon brush during use by analyzing the contact force, temperature distribution and contact duration between the carbon brush and the busbar, and generates the remaining service life of the carbon brush; The feature extraction module is used to extract joint feature data from the temperature field data, stress field data and wear rate prediction data. The feature extraction module uses a principal component analysis algorithm to reduce the dimension of multidimensional data and compress redundant feature information into a low-dimensional feature vector.

[0044] The multi-dimensional thermal-mechanical coupling module meshes the contact area between the carbon brush and the busbar based on the finite element method (FEM), and calculates and generates temperature field data and stress field data through the interaction between the temperature field and the stress field.

[0045] Specifically, the finite element method divides the contact area into multiple small grid units, establishes a mathematical model for each grid unit, and calculates the temperature distribution and stress distribution of each grid unit. The friction between the carbon brush and the busbar will generate heat at the contact point, which will cause the temperature of the contact area to rise. The temperature field data is used to describe the distribution of this heat. The stress field data is used to describe the contact force in the contact area between the carbon brush and the busbar, especially the local pressure and stress distribution between the carbon brush and the busbar. The stress field data not only helps to describe the mechanical load in the contact area, but also provides key mechanical parameters for further calculation of the wear rate.

[0046] In the present invention, the module first uses the fused data (including carbon brush height, temperature, vibration, etc.) from the data preprocessing unit to establish a model, and then applies the finite element method to finely model the contact area, thereby obtaining high-precision temperature field and stress field data. In particular, the temperature field data can provide a theoretical basis for subsequent carbon brush wear and equipment overheating warnings; while the stress field data provides a detailed description of the pressure distribution at the contact point between the carbon brush and the busbar, which helps to identify the risk of wear or damage caused by excessive contact pressure.

[0047] The multi-dimensional thermal-mechanical coupling module can obtain accurate temperature distribution and contact stress distribution by accurately calculating the thermal-mechanical interaction between the carbon brush and the busbar, providing necessary data support for subsequent wear prediction and risk assessment. Through this module, the system can simulate and predict the risks of overheating areas, abnormal contact force, etc. that may occur during the contact between the carbon brush and the busbar, and provide early warning in time, thereby effectively avoiding equipment damage.

[0048] The wear rate prediction module calculates the wear rate of carbon brushes based on Archard's law. Archard's law is an empirical formula commonly used to describe the wear rate, especially for the contact behavior of materials in the process of friction and wear. Its basic formula is:

[0049] in, is the wear amount (unit: mm³), is the wear coefficient (dimensionless), is the contact force (unit: N), is the sliding distance (unit: mm).

[0050] In the present invention, the wear rate prediction module takes the temperature field data and the stress field data as input to calculate the wear rate of the carbon brush. By obtaining the contact force between the carbon brush and the busbar ( ) and the sliding distance generated during contact ( ), the wear of the carbon brush can be predicted based on Archard's law. Wear coefficient is a constant determined based on experimental data and is usually related to factors such as material properties and friction conditions. It will be dynamically adjusted according to different working conditions and carbon brush material characteristics.

[0051] Through the wear rate prediction module, the system can calculate the wear rate of the carbon brush and estimate the remaining service life of the carbon brush based on the wear rate. The predicted wear rate can not only reflect the health status of the carbon brush in real time, but also provide early warning information to maintenance personnel by monitoring the progress of wear, so as to arrange maintenance or replacement of carbon brushes in advance, thereby avoiding system failure or equipment downtime caused by excessive wear.

[0052] The main function of the feature extraction module is to extract joint feature data from temperature field data, stress field data and wear rate prediction data for further risk assessment by subsequent analysis units. In order to reduce the dimension of data and improve analysis efficiency, the feature extraction module adopts Principal Component Analysis (PCA) The algorithm performs dimensionality reduction.

[0053] Principal component analysis (PCA) is a commonly used dimensionality reduction technique that converts multidimensional data into a few principal components through linear transformation, thereby reducing redundant data and retaining most of the information. The core idea of ​​the PCA algorithm is to find the principal components in the data (i.e., the direction of maximum variance), and then project the original data onto these principal components to form new low-dimensional data. Through PCA, the feature extraction module can convert multidimensional data such as temperature, stress, and wear into a few feature vectors, which can effectively describe the working status and potential risks of carbon brushes and busbars.

[0054] Through PCA dimensionality reduction, the feature extraction module significantly reduces the complexity of the input data and merges multiple variables into a few effective features. This not only improves the computational efficiency, but also enhances the accuracy of subsequent risk assessment and prediction. The reduced feature data focuses more on the key information in the contact state between the carbon brush and the busbar, avoiding the interference of redundant data on the analysis results. In addition, the joint feature data generated by the feature extraction module provides efficient input for subsequent nonlinear analysis, ensuring the real-time and accuracy of risk assessment.

[0055] The modeling unit completes the whole process from data fusion to risk assessment through its three core modules: multi-dimensional thermal-mechanical coupling module, wear rate prediction module and feature extraction module. The multi-dimensional thermal-mechanical coupling module accurately calculates the temperature field and stress field, providing a scientific basis for wear prediction and fault assessment; the wear rate prediction module calculates the wear rate of the carbon brush based on Archard's law and combines temperature and stress data to help estimate the remaining service life of the carbon brush; the feature extraction module reduces the dimension of the data through principal component analysis to ensure the efficient extraction of key features in the contact state between the carbon brush and the busbar, providing accurate data support for subsequent risk assessment and protection measures. The synergistic effect of these modules enables the system to accurately monitor the working status of the carbon brush, promptly detect potential problems and take corresponding protection measures, thereby improving the operating efficiency and reliability of the oil-to-electric carbon brush system.

[0056] The analysis unit is used to perform nonlinear analysis on the joint feature data to generate risk scoring data, and perform risk assessment on the operating status of the carbon brush and the busbar according to the risk scoring data, thereby triggering an early warning or executing protective measures.

[0057] Preferably, Figure 4 As shown, the analysis unit includes: A nonlinear analysis module, used for performing nonlinear analysis on the joint feature data, wherein the nonlinear analysis module is based on a multi-layer perceptron neural network, uses a supervised learning method to train the joint feature data, and generates risk score data through nonlinear mapping, wherein the risk score data represents the potential failure risk of the carbon brush and busbar system, and the neural network includes an input layer, multiple hidden layers and an output layer, and uses a back propagation algorithm to update weights and minimize errors; The risk assessment module is used to perform risk assessment on the operating status of the carbon brush and the busbar according to the risk scoring data. The risk assessment module determines the risk level of the carbon brush and the busbar according to the set threshold value. The risk level includes low risk, medium risk and high risk. The risk assessment result determines whether to trigger an early warning or start protective measures.

[0058] The core of the nonlinear analysis module is training and prediction based on the multilayer perceptron (MLP) neural network. The multilayer perceptron is a feedforward neural network that is usually used to process complex data with nonlinear relationships. The structure of the neural network usually includes an input layer, multiple hidden layers, and an output layer, where the nonlinear activation function of the hidden layer enables the network to learn and fit complex nonlinear mappings.

[0059] In the present invention, the input layer receives the joint feature data generated from the modeling unit, which includes temperature field data, stress field data and wear rate prediction data. Through the multi-layer structure of the multi-layer neural network, the data is processed and converted multiple times by the hidden layer, and finally outputs a risk score data representing the potential failure risk of the carbon brush and busbar system.

[0060] Specifically, supervised learning means that during the training process, the network will adjust the network weights through labeled historical data (i.e. known fault and non-fault data) so that the final output risk score can accurately reflect the health status of the carbon brushes and busbars. The back propagation algorithm is used during the training process, which calculates the output error and propagates it back to each layer of the network, updating the network weights layer by layer until the network can minimize the prediction error. Each round of updates enhances the network's ability to predict risk scores.

[0061] By performing nonlinear analysis on joint feature data through neural networks, the system can accurately map multi-dimensional, complex physical features (such as temperature, stress, wear rate, etc.) to a concise risk score data. This nonlinear mapping capability enables the system to identify and capture the complex interactive relationship between carbon brushes and busbars, accurately reflecting their potential failure risks. Through the training of multi-layer perceptron neural networks, the analysis module can continuously improve the accuracy of risk assessment, thereby improving the early warning capability of the entire system.

[0062] The main task of the risk assessment module is to conduct a quantitative risk assessment of the operating status of carbon brushes and busbars based on the risk score data generated by the nonlinear analysis module. The module classifies the risk scores according to the preset risk threshold, determines the risk level of carbon brushes and busbars, and triggers corresponding warnings or protection measures based on the risk level.

[0063] Specifically, the risk score is usually a value ranging from 0 to 100, indicating the health of the carbon brushes and busbars. The module compares them with the preset risk threshold and divides them into the following risk levels: Low Risk: A risk score below 30 indicates that the equipment is in good condition and does not have any potential failures.

[0064] Medium risk: A risk score between 30 and 70 indicates that the equipment has some risk of failure and requires monitoring and possible maintenance.

[0065] High risk: A risk score above 70 indicates that the equipment is at risk of serious failure and protective measures must be taken immediately to prevent damage to the equipment.

[0066] These thresholds are the result of debugging and optimization based on historical data, equipment performance, working environment, etc. In actual applications, the risk assessment module dynamically evaluates the working status of the equipment according to the changes in risk scores and provides corresponding action plans for each risk level, such as whether to trigger an early warning or execute automatic protection measures.

[0067] The risk assessment module compares the risk score data with the preset threshold to quickly determine the health status of the equipment and respond in real time. When the risk score is in the high-risk range, the system will immediately trigger protection measures to prevent damage or downtime of the equipment due to failure; when the score is medium risk, the system will issue an early warning prompt to remind maintenance personnel to check or make appropriate adjustments; in the case of low risk, the system will continue to operate normally to reduce unnecessary maintenance operations.

[0068] This dynamic risk assessment mechanism greatly improves the system's response speed, enabling the system to promptly detect potential problems and take corresponding measures, thereby effectively ensuring the long-term stable operation of the equipment.

[0069] In the example, suppose that during the operation of the system, the nonlinear analysis module obtains a risk score of 65 through neural network analysis, indicating that there is a medium risk of contact between the carbon brush and the busbar. The risk assessment module compares the score with the preset medium risk threshold, confirms that the risk level is medium risk, and the system triggers an early warning to prompt maintenance personnel to check the equipment. At this time, maintenance personnel can conduct targeted inspections based on the real-time displayed data (such as temperature, vibration, etc.) to prevent the problem from further deteriorating.

[0070] If the risk score is 85, the system will identify it as a high risk and immediately trigger automatic shutdown protection, connecting the control system to the power supply and cutting off the power supply to prevent further damage to the equipment due to the fault.

[0071] Through the close collaboration of the nonlinear analysis module and the risk assessment module, the system can not only monitor the working status of the carbon brushes and busbars in real time, but also accurately assess the potential risk of failure and respond promptly when the risk occurs. This mechanism ensures that the oil-to-electric carbon brush height identification and early warning system can always maintain high stability and reliability when operating in complex environments, while reducing the risk of equipment damage and downtime, providing maintenance personnel with efficient support tools to ensure that the equipment can work continuously, safely and efficiently.

[0072] Preferably, Figure 5 As shown, when the risk level is high, the risk assessment module triggers protection measures, including: Automatic shutdown protection, used to stop the operation of carbon brushes or busbar equipment when high risk is detected. The automatic shutdown protection is connected to the equipment power supply through the control system to interrupt the power supply of the equipment in real time to prevent further damage to the equipment; Adjusting the load distribution is used to reduce the risk of failure by controlling the load distribution of the busbar when the risk assessment is medium risk. The load distribution is adjusted by adjusting the load balancing module to ensure reasonable distribution of the busbar load and reduce equipment wear.

[0073] Automatic shutdown protection is the most direct and effective protection measure taken when the system evaluates that the carbon brush or busbar equipment is in a high-risk state. Its working principle is based on the high-risk score data generated by the risk assessment module. When the risk score exceeds the predetermined threshold (such as 70), it means that the status of the equipment has entered the potential risk range of serious failure. In order to prevent the equipment from causing more serious failures due to overload, overheating or other factors, the system will immediately perform an automatic shutdown operation.

[0074] Automatic shutdown protection is achieved by connecting the control system to the power supply of the equipment. During this process, the control system cuts off the power supply of the equipment or stops the operation of the equipment through signal instructions. This power cut-off process is completed within milliseconds to ensure the rapidity and accuracy of the equipment shutdown, thereby reducing further damage caused by the fault. For example, when the carbon brush is severely worn or the temperature of the busbar is too high, the automatic shutdown protection will cut off the power supply to prevent equipment damage or safety accidents.

[0075] Through automatic shutdown protection, when the equipment enters a high-risk state, the system can immediately stop the operation of the equipment to avoid further damage caused by carbon brush wear, overheating, excessive vibration, etc. This not only improves the safety of the equipment, but also effectively avoids economic losses caused by equipment damage. In practical applications, automatic shutdown protection can ensure the long-term stable operation of the equipment and prevent large-scale shutdowns caused by system failures.

[0076] In the example, suppose that during a certain operation, the risk assessment module detects that the contact force between the carbon brush and the busbar is too large, resulting in a system risk score of 80 (high risk). At this time, the system will immediately trigger automatic shutdown protection, instructing the control system to cut off the power supply and stop the operation of the carbon brush or busbar equipment. This can avoid damage to the equipment due to overload, such as motor burnout, control system damage, and other problems.

[0077] Adjusting the load distribution is a protective measure taken in a medium-risk state. When the risk assessment module detects that the equipment is in a medium-risk state (for example, the risk score is between 30 and 70), although the equipment does not need to be shut down immediately, there is still a certain risk of failure. In this case, the system adjusts the load distribution to reduce the possibility of equipment failure and reduce excessive wear of the equipment, thereby extending the service life of the equipment and reducing the probability of failure.

[0078] The principle of adjusting the load distribution is to redistribute the load of the busbar by controlling the load balancing module of the busbar. This process adjusts the load distribution strategy in real time based on the current load conditions of the equipment (such as the contact pressure, temperature, vibration and other data between the carbon brush and the busbar) to ensure that the load distribution in each busbar area is more uniform and avoid local overload. The optimization of load distribution can be achieved by adjusting the pressure state of the carbon brush, adjusting the current distribution or by adjusting the mechanical settings in the system.

[0079] For example, a part of the busbar may be under great pressure due to excessive use or severe wear. At this time, adjusting the load distribution module will reduce the load of this part through the control system and transfer the load to other parts, thereby reducing excessive wear and extending the service life of the equipment.

[0080] By adjusting the load distribution, the system can reduce the load on a certain part of the equipment and avoid equipment wear or damage caused by local overload. Especially in medium-risk conditions, adjusting the load distribution can effectively reduce the probability of equipment failure and ensure stable operation of the equipment. For example, when uneven wear occurs between the carbon brush and the busbar, the system can adjust the load in real time to reduce local overload and prevent serious equipment failure without effective warning.

[0081] Example, suppose that during a certain monitoring phase, the system detects abnormal temperature and vibration data of a certain part of the busbar, indicating that the area is heavily loaded and the risk score reaches 50 (medium risk). The system adjusts the load balancing module to redistribute the load of the busbar, distributing the load of the area with higher pressure to the area with lower pressure to reduce the burden of that part and ensure that the equipment is not damaged due to local overload.

[0082] Preferably, the system further comprises: A storage unit, used to store the carbon brush height data, carbon brush temperature data, carbon brush vibration data, busbar surface image data and environmental parameter data, the storage unit adopts a high-performance hard disk or solid-state hard disk and supports real-time data backup; The incremental learning module is used to optimize the computing power of the modeling unit and the analysis unit based on historical data and real-time data. The incremental learning module updates the model through regular training, so that the system can continuously improve the prediction accuracy according to new data.

[0083] The storage unit is responsible for storing various data collected in the system, including carbon brush height data, carbon brush temperature data, carbon brush vibration data, busbar surface image data, and environmental parameter data. After these data are collected in real time by various sensors, they need to be stored efficiently and securely for subsequent analysis, modeling, and decision-making. In order to meet the needs of large data volumes and high-frequency data storage, the storage unit uses a high-performance hard disk (HDD) or solid-state drive (SSD), and supports real-time data backup to ensure data efficiency and reliability.

[0084] As mainstream data storage devices, high-performance hard disk drives (HDDs) and solid-state drives (SSDs) have high read and write speeds and stable performance. SSDs are particularly suitable for storing frequently written and read data, such as carbon brush temperature, vibration, and image data. These data have high real-time requirements and may accumulate over time to form a large amount of data. Therefore, the high-speed read and write capabilities of SSDs ensure that data can be quickly stored and retrieved at any time.

[0085] Real-time data backup uses data redundancy strategies to ensure that important data is not lost even if hardware failure or system crash occurs. The backup mechanism automatically synchronizes and copies data to another storage medium or cloud at regular intervals to facilitate data recovery and prevent single point failures.

[0086] By combining high-performance hard disks and solid-state drives, the storage unit can ensure that a large amount of frequently generated data in the system can be quickly and securely stored and accessed. In addition, real-time data backup ensures data security and avoids data loss due to storage medium failure. This provides solid data protection for subsequent data processing, modeling, analysis and risk assessment.

[0087] In the embodiment, during operation, the carbon brush temperature data and busbar image data will be stored in real time in the SSD. At the same time, the system will regularly back up this data to a remote storage device to prevent data loss due to hardware failure. The real-time backup mechanism of the storage unit ensures that all critical data can be restored through backup even in the event of a failure, thereby ensuring the continuity and stability of the system.

[0088] The core task of the incremental learning module is to optimize the computing power of the modeling unit and analysis unit in the system through continuous learning and training of historical data and real-time data. Incremental learning is a learning method that can continuously adjust existing models based on new data. Unlike traditional batch learning methods, incremental learning does not require retraining the entire model every time new data is added. Instead, it enables the model to adapt to changing input data and environmental conditions at any time through incremental updates of new data.

[0089] In the present invention, the incremental learning module regularly receives historical data and newly collected real-time data in the storage unit, and optimizes the existing modeling unit (responsible for the calculation of temperature field, stress field and wear rate) and analysis unit (responsible for risk scoring and evaluation) by comparing and analyzing the differences between historical data and current data. The incremental learning algorithm adjusts the parameters in the modeling process so that the model can gradually improve the prediction accuracy, especially in predicting carbon brush wear, failure risk, etc.

[0090] The workflow of the incremental learning module is as follows: Data input: New data enters the system through sensor information collected in real time, and the storage unit provides this data to the incremental learning module.

[0091] Model update: The incremental learning module continuously adjusts and optimizes the modeling and analysis algorithms by comparing new data with historical data. For example, by learning new wear patterns or carbon brush temperature changes under different operating conditions, the model's ability to predict carbon brush life can be optimized.

[0092] Real-time feedback: The updated model is tested in real time on new data, and the optimization results are fed back to the modeling and analysis units in the system, thereby improving the accuracy of the system's prediction of carbon brush and busbar faults.

[0093] The incremental learning module can make the system better adapt to factors such as environmental changes, working conditions and equipment aging by continuously optimizing and updating the system's modeling and analysis capabilities. As the system continues to accumulate historical data, incremental learning can continuously improve the model's prediction accuracy and gradually reduce errors, especially in terms of carbon brush wear, failure risk and equipment life prediction. Systems that use incremental learning for a long time will be more accurate and reliable, and can reflect the true status of the equipment in real time.

[0094] In the embodiment, it is assumed that the system gradually learns the temperature change pattern of the carbon brush through the incremental learning algorithm during operation. When the temperature data of the carbon brush under a specific working condition is abnormal, the incremental learning module can identify and adjust the prediction model in time, so as to more accurately predict the remaining service life of the carbon brush, and issue an early warning to notify maintenance personnel to replace it. As the system continues to accumulate temperature data under different working conditions, the incremental learning module can quickly adapt to new working conditions, further improving the system's prediction accuracy for the equipment.

[0095] Preferably, the data acquisition unit is connected to the data preprocessing unit, the modeling unit and the analysis unit via a communication network. The communication network uses an industrial-grade Ethernet protocol, is capable of transmitting the collected real-time data to each unit, and returns the processing results to the user or the control system. The transmission delay of the communication network does not exceed 10 milliseconds.

[0096] Industrial Ethernet protocol is a communication protocol widely used in industrial control and automation systems. It is strengthened and optimized based on the standard Ethernet protocol (such as IEEE 802.3) to ensure high stability and reliability in harsh environments (such as high temperature, high humidity, strong electromagnetic interference, etc.).

[0097] In the present invention, the communication network is responsible for real-time transmission of data from different modules (such as data acquisition, data preprocessing, modeling and analysis). The specific implementation principle is as follows: Each module is connected via industrial Ethernet and can transmit the collected real-time data (such as carbon brush temperature, vibration, image data, etc.) to other modules. For example, the data acquisition unit is responsible for collecting carbon brush temperature and busbar image data in real time, and then transmits it to the data preprocessing unit via Ethernet protocol for further processing.

[0098] In order to ensure that the system can perform real-time monitoring and early warning, the communication network must have the characteristics of low latency and high transmission rate. Through the industrial Ethernet protocol, the delay of data transmission can be controlled within 10 milliseconds, which is crucial for systems with extremely high real-time requirements. Low latency ensures fast response of data between modules, allowing the system to process and feedback the operating status of carbon brushes and busbars in a timely manner.

[0099] The results of data processing and modeling analysis need to be promptly returned to the user or control system for further decision-making or execution. In this system, after the analysis unit completes the risk assessment, it will transmit the risk score data and warning information back to the control system or display unit through the same Ethernet protocol, thereby achieving real-time decision-making and operation.

[0100] The communication network is not only used for data transmission, but also for feedback of control instructions. For example, when the system triggers a protective measure (such as automatic shutdown) based on the risk assessment results, the control signal will be transmitted to the execution module through the network to control the carbon brush or busbar equipment to shut down. Real-time two-way communication enables the entire system to dynamically respond to changing operating environments and equipment status.

[0101] To ensure efficient data transmission and low latency, the system uses industrial-grade Ethernet switches and dedicated network connections to ensure stable connections even in complex environments. The specific implementation steps include: (1) The system adopts a star topology, and all modules are connected through an Ethernet switch to ensure direct communication between each module. At the same time, the switch uses industrial-grade equipment that can withstand high-load, high-frequency data transmission requirements and has anti-interference capabilities.

[0102] (2) In terms of communication protocol selection, the system adopts the Gigabit Ethernet standard with a transmission rate of 1Gbps, which can meet the high-speed transmission requirements of carbon brush and busbar related data. Most of the real-time collected data, such as temperature and vibration, are small in size, but image data and modeling data such as temperature field and stress field require larger bandwidth and higher transmission speed, so a high-speed network is necessary.

[0103] (3) To ensure real-time performance, the system optimizes data transmission by selecting low-latency, low-jitter industrial Ethernet devices and combining them with priority queuing technology (QoS) to ensure that high-priority data (such as sensor data and warning signals) can pass through the network first and ensure timely transmission.

[0104] (4) In order to improve the reliability of the system, the network connection adopts a redundant design. For example, through dual-link redundancy technology, if one connection line fails, the other line can immediately take over the data transmission, ensuring the high availability of the system.

[0105] The low latency of the communication network ensures that data between modules can be transmitted quickly, thus achieving real-time data processing and rapid response. This is particularly important for the oil-to-electric carbon brush height identification and warning system, especially when the contact state between the carbon brush and the busbar changes, the system can obtain data and respond in a very short time. The system is connected through a high-speed industrial Ethernet connection, ensuring that large amounts of data (such as image data and modeling results) can flow efficiently. Even when the equipment is busy, the data transmission can remain stable and high-speed, avoiding data delays or losses caused by network bottlenecks. The anti-interference and anti-harsh environment capabilities of the industrial Ethernet protocol ensure the long-term stable operation of the system in an industrial environment. Even in an environment with electromagnetic interference and large changes in temperature and humidity, the system can ensure the stability and accuracy of data transmission. The real-time two-way communication mechanism enables the control system to make decisions based on the feedback from the analysis unit and to execute protection measures in a timely manner through network instructions. Whether it is automatic shutdown protection or load adjustment, the communication network can ensure the rapid and accurate execution of decisions.

[0106] In this example, if the system detects that the carbon brush wear exceeds the limit, a high risk score (risk score of 80) is generated, and the analysis unit sends an early warning signal. Through the communication network, the risk score data will be transmitted to the control system in real time, and the automatic shutdown protection will be triggered to cut off the power supply. The entire process from detection to shutdown completion takes no more than 10 milliseconds, ensuring that the system responds quickly to prevent further equipment damage.

[0107] Preferably, the system further comprises: The display unit is used to display the current status, risk level and processing suggestions of the carbon brush and the busbar. The display unit adopts a touch screen interface and supports a real-time refresh function for reference and operation by maintenance personnel.

[0108] The design of the display unit is based on the touch screen display technology widely used in modern industry, combining the needs of real-time data update and user interaction. The main function of the display unit is to show the equipment status monitored in the system to maintenance personnel in real time, and to help users understand the health status and potential risks of carbon brushes and busbars through a graphical interface. The system transmits data from various modules (such as data acquisition, modeling, analysis, etc.) to the display unit through a communication network, thus forming a real-time, intuitive visual interface.

[0109] The display unit uses a touch screen interface as an interactive mode for input and output, supporting user operation through touch. The touch screen interface enables quick query of system data, setting adjustment, and input of risk handling instructions. For example, maintenance personnel can view carbon brush temperature, busbar image, and real-time risk score through the touch screen. The touch interface can also be used to start the system's automatic protection measures or perform manual control.

[0110] The display unit can receive the latest data from the data acquisition unit and analysis unit in real time, and update the display content in real time. The data acquisition and analysis modules in the system regularly generate new data, such as carbon brush temperature, vibration and other information, or risk score data. These data are transmitted to the display unit through the communication network, triggering dynamic updates of the interface, thereby ensuring that maintenance personnel always view the latest status of the equipment. This real-time refresh function is the key to ensuring that the system can respond immediately to changes.

[0111] Information display: The data displayed by the display unit includes: The current status of carbon brushes and busbars: including equipment temperature, vibration, wear and other indicators, displayed in the form of numbers and charts to help maintenance personnel intuitively understand the equipment operation status.

[0112] Risk level: Based on the risk assessment results of the analysis unit, the risk status of carbon brushes and busbars is displayed in real time. The system automatically displays the current risk level (low risk, medium risk, high risk) based on the risk score, and may enhance the visual representation of the risk through color coding (e.g. green, yellow, red).

[0113] Treatment suggestions: Based on the risk level and the results of system analysis, the display unit will provide maintenance personnel with treatment suggestions or early warning information. For example, when the risk is high, the display unit may suggest "immediate shutdown" or "check carbon brush wear", and when the risk is low, the display unit will display "equipment is operating normally, continue monitoring".

[0114] Operation and interaction: In addition to displaying information, the touch screen also allows maintenance personnel to interact directly with the system. For example, maintenance personnel can choose to view historical data, adjust equipment operating parameters, or manually trigger early warning and protection measures. This interactive method greatly improves the convenience and real-time nature of operation and avoids the complex operation of traditional button-type interfaces.

[0115] The display unit uses an industrial-grade touch screen display, combined with an embedded control system to display and process data. Its implementation involves the following key technologies: Touch screen display: The display unit uses a high-resolution industrial touch screen to ensure clear and delicate display, capable of displaying the device's temperature, vibration data, image data, etc. Through touch operation, users can click on different data modules to view more detailed information. The touch screen also supports multi-touch, making it convenient for users to perform multiple operations at the same time.

[0116] Real-time data refresh: The display unit exchanges real-time data with the data acquisition, analysis and modeling units to ensure that the information display is always synchronized with the system status. The system uses industrial Ethernet protocols (such as Gigabit Ethernet) for high-speed, low-latency data transmission, so that the information on the screen can be consistent with the system status, and the transmission delay is usually controlled within 10 milliseconds.

[0117] Data visualization and graphical display: By integrating chart and image display software, the display unit can convert the collected data (such as temperature, vibration, etc.) into visual charts or real-time images. These graphical data displays help maintenance personnel quickly identify abnormalities, such as temperature change curves, depth images of carbon brush wear, etc.

[0118] System configuration and data interaction: The embedded control unit receives user instructions through the touch screen and updates the system configuration in real time. For example, maintenance personnel can adjust equipment operating parameters (such as load, temperature threshold, etc.) through the touch screen and choose whether to enable specific protection measures. In addition, the display unit can also query data, display historical data or conduct detailed analysis of equipment status according to user needs.

[0119] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems or computer program products. Therefore, the present application may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware.

[0120] The above are only embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the claims of the present application.

Claims

1. A oil-to-electric carbon brush height identification and warning system, characterized in that: include: A data acquisition unit, used to collect carbon brush height data, temperature data, vibration data, busbar surface image data and environmental parameter data; A data preprocessing unit, used for preprocessing the collected data to generate fused data; A modeling unit, for calculating and generating temperature field data, stress field data and wear rate prediction data based on the fused data by adopting a multi-dimensional thermo-mechanical coupling modeling method, and extracting joint feature data; The analysis unit is used to perform nonlinear analysis on the joint feature data to generate risk scoring data, and perform risk assessment on the operating status of the carbon brush and the busbar according to the risk scoring data, thereby triggering an early warning or executing protective measures.

2. The oil-to-electric carbon brush height identification and warning system according to claim 1 is characterized in that: The data acquisition unit comprises: Laser distance sensor, used to collect the height data of carbon brush; Infrared thermal imager, used to collect temperature data of carbon brushes; High-frequency vibration sensor, used to collect vibration data of carbon brushes; Visual depth camera, used to collect busbar surface image data; Humidity and air pressure sensors are used to collect environmental parameter data.

3. The oil-to-electric carbon brush height identification and warning system according to claim 1 is characterized in that: The data preprocessing unit comprises: A data formatting module, used to convert the carbon brush height data, carbon brush temperature data, carbon brush vibration data, busbar surface image data and environmental parameter data collected from different sensors into a unified format; Noise filtering module, used to remove noise from the collected data to ensure the accuracy of the data; The time alignment module is used to time align the carbon brush height data, carbon brush temperature data, carbon brush vibration data, busbar surface image data and environmental parameter data collected by different sensors to generate fused data.

4. The oil-to-electric carbon brush height identification and warning system according to claim 1 is characterized in that: The modeling unit comprises: A multi-dimensional thermal-mechanical coupling module, which is used to mesh the contact area between the carbon brush and the busbar using the finite element method based on the fusion data, and calculate and generate temperature field data and stress field data according to the interaction between the temperature field and the stress field in the contact area between the carbon brush and the busbar. The temperature field data is used to describe the heat distribution generated during the friction process, and the stress field data is used to describe the contact force between the carbon brush and the busbar. The stress field data is further used to evaluate the pressure distribution of the contact point; A wear rate prediction module is used to calculate the wear rate of the carbon brush according to the temperature field data and the stress field data. The wear rate prediction module is based on Archard's law and predicts the wear rate of the carbon brush during use by analyzing the contact force, temperature distribution and contact duration between the carbon brush and the busbar, and generates the remaining service life of the carbon brush; The feature extraction module is used to extract joint feature data from the temperature field data, stress field data and wear rate prediction data. The feature extraction module uses a principal component analysis algorithm to reduce the dimension of multidimensional data and compress redundant feature information into a low-dimensional feature vector.

5. The oil-to-electric carbon brush height identification and warning system according to claim 1 is characterized in that: The analysis unit comprises: A nonlinear analysis module, used for performing nonlinear analysis on the joint feature data, wherein the nonlinear analysis module is based on a multi-layer perceptron neural network, uses a supervised learning method to train the joint feature data, and generates risk score data through nonlinear mapping, wherein the risk score data represents the potential failure risk of the carbon brush and busbar system, and the neural network includes an input layer, multiple hidden layers and an output layer, and uses a back propagation algorithm to update weights and minimize errors; The risk assessment module is used to perform risk assessment on the operating status of the carbon brush and the busbar according to the risk scoring data. The risk assessment module determines the risk level of the carbon brush and the busbar according to the set threshold value. The risk level includes low risk, medium risk and high risk. The risk assessment result determines whether to trigger an early warning or start protective measures.

6. The oil-to-electric carbon brush height identification and warning system according to claim 5 is characterized in that: When the risk level is high, the risk assessment module triggers protective measures, including: Automatic shutdown protection, used to stop the operation of carbon brushes or busbar equipment when high risk is detected. The automatic shutdown protection is connected to the equipment power supply through the control system to interrupt the power supply of the equipment in real time to prevent further damage to the equipment; Adjusting the load distribution is used to reduce the risk of failure by controlling the load distribution of the busbar when the risk assessment is medium risk. The load distribution is adjusted by adjusting the load balancing module to ensure reasonable distribution of the busbar load and reduce equipment wear.

7. The oil-to-electric carbon brush height identification and warning system according to claim 1 is characterized in that: The system further comprises: A storage unit, used to store the carbon brush height data, carbon brush temperature data, carbon brush vibration data, busbar surface image data and environmental parameter data, the storage unit adopts a high-performance hard disk or solid-state hard disk and supports real-time data backup; The incremental learning module is used to optimize the computing power of the modeling unit and the analysis unit based on historical data and real-time data. The incremental learning module updates the model through regular training, so that the system can continuously improve the prediction accuracy according to new data.

8. The oil-to-electric carbon brush height identification and warning system according to claim 1 is characterized in that: The data acquisition unit is connected to the data preprocessing unit, the modeling unit and the analysis unit through a communication network. The communication network uses an industrial-grade Ethernet protocol, can transmit the collected real-time data to each unit, and return the processing results to the user or the control system. The transmission delay of the communication network does not exceed 10 milliseconds.

9. The oil-to-electric carbon brush height identification and warning system according to claim 1 is characterized in that: The system further comprises: The display unit is used to display the current status, risk level and processing suggestions of the carbon brush and the busbar. The display unit adopts a touch screen interface and supports a real-time refresh function for reference and operation by maintenance personnel.

Images