CN111539374B - Fault diagnosis method of rail train bearing based on multi-dimensional data space - Google Patents
Fault diagnosis method of rail train bearing based on multi-dimensional data spaceDownload PDFInfo
- Publication number
- CN111539374B CN111539374BCN202010384367.3ACN202010384367ACN111539374BCN 111539374 BCN111539374 BCN 111539374BCN 202010384367 ACN202010384367 ACN 202010384367ACN 111539374 BCN111539374 BCN 111539374B
- Authority
- CN
- China
- Prior art keywords
- data
- train
- sample
- dimensional
- dimensional data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/02—Preprocessing
- G—PHYSICS
- G01—MEASURING; TESTING
- G01M—TESTING STATIC OR DYNAMIC BALANCE OF MACHINES OR STRUCTURES; TESTING OF STRUCTURES OR APPARATUS, NOT OTHERWISE PROVIDED FOR
- G01M13/00—Testing of machine parts
- G01M13/04—Bearings
- G01M13/045—Acoustic or vibration analysis
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/12—Classification; Matching
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Signal Processing (AREA)
- Acoustics & Sound (AREA)
- Train Traffic Observation, Control, And Security (AREA)
Abstract
Translated fromChinese
Description
Translated fromChinese技术领域technical field
本发明涉及轨道列车轴承故障诊断领域，尤其是涉及一种基于多维数据空间的轨道列车轴承故障诊断方法。The invention relates to the field of rail train bearing fault diagnosis, in particular to a rail train bearing fault diagnosis method based on multi-dimensional data space.
背景技术Background technique
随着我国城市轨道交通的跨越式高速发展，轨道交通面临着严峻的安全问题。轨道车辆具有运行状态复杂多变、噪声干扰大、故障模式复杂、可靠性要求高等特点，对故障诊断提出了较高的要求，滚动轴承是高速列车中最易损坏的重要零部件之一，工作环境复杂，当受到过大的载荷冲击或者安装设计不当、润滑状态不良等因素影响时，都有可能出现故障，造成重大安全隐患和巨大经济损失，因此快速、准确地对列车轴承进行故障诊断，及时定位故障，在确保轨道交通系统安全、可靠、稳定运行中具有重要意义。With the rapid development of urban rail transit in my country, rail transit is facing severe safety problems. Rail vehicles have the characteristics of complex and changeable operating states, large noise interference, complex failure modes, and high reliability requirements, which put forward higher requirements for fault diagnosis. Rolling bearings are one of the most vulnerable and important components in high-speed trains. The working environment It is complicated. When it is affected by factors such as excessive load impact or improper installation design, poor lubrication state, etc., failure may occur, resulting in major safety hazards and huge economic losses. Therefore, the fault diagnosis of train bearings can be carried out quickly and accurately. Locating faults is of great significance in ensuring the safe, reliable and stable operation of rail transit systems.
目前，轴承的故障诊断主要以振动信号分析为基础，近十年来已经发展了丰富的轴承诊断技术，例如包络谱分析、小波分析、经验模式分解等，这些方法通过不断优化，已经在工程应用领域发挥了一定的作用，准确率不断提升。但是仍然存在一定局限性，包括：采集的信息种类较为局限，对车辆的实际运行状态、不同车辆本身的特点考虑的不够全面，存在大量的有效信息浪费的情况，缺乏普遍适用性；采用单一维度的物理信号很难适应复杂的工况，陷入局部诊断的误区，诊断可靠性不足，易造成安全隐患；对于早期故障的诊断经验不足，诊断实时性还有待提高。At present, the fault diagnosis of bearings is mainly based on the analysis of vibration signals. In the past ten years, a wealth of bearing diagnosis techniques have been developed, such as envelope spectrum analysis, wavelet analysis, empirical mode decomposition, etc. These methods have been applied in engineering through continuous optimization. The field has played a certain role, and the accuracy rate has been continuously improved. However, there are still some limitations, including: the types of information collected are relatively limited, the actual operating state of the vehicle and the characteristics of different vehicles are not considered comprehensively, there is a lot of waste of effective information, and the lack of universal applicability; a single dimension is adopted. It is difficult to adapt to the complex working conditions of the physical signal of the fault, and it falls into the misunderstanding of local diagnosis. The reliability of the diagnosis is insufficient, and it is easy to cause potential safety hazards; the diagnosis of early faults is insufficient, and the real-time diagnosis needs to be improved.
发明内容SUMMARY OF THE INVENTION
本发明的目的就是为了克服上述现有技术存在的缺陷而提供一种基于多维数据空间的轨道列车轴承故障诊断方法。The purpose of the present invention is to provide a fault diagnosis method for rail train bearings based on multi-dimensional data space in order to overcome the above-mentioned defects in the prior art.
本发明的目的可以通过以下技术方案来实现：The object of the present invention can be realized through the following technical solutions:
一种基于多维数据空间的轨道列车轴承故障诊断系统，该系统包括：A rail train bearing fault diagnosis system based on multi-dimensional data space, the system includes:
数据采集单元：用以采集轴承检测信号和列车状态识别辅助信息数据；Data acquisition unit: used to collect bearing detection signals and auxiliary information data for train status identification;
数据发送单元：用以将数据采集单元采集到的数据发送至运行状态识别单元；Data sending unit: used to send the data collected by the data collection unit to the running state identification unit;
运行状态识别单元：接收数据发送单元传送的数据，并对采集到的数据进行预处理，通过分类识别获得列车的运行状态；Running state identification unit: receives the data transmitted by the data sending unit, preprocesses the collected data, and obtains the running state of the train through classification and identification;
计算单元：用以计算质心和中心值，以及各检测样本数据点与正常样本的质心或中心之间的距离或差值，并将结果传送至结果显示单元；Calculation unit: used to calculate the centroid and center value, as well as the distance or difference between each test sample data point and the centroid or center of the normal sample, and transmit the results to the result display unit;
结果显示单元：用以进行阈值判断，将在正常阈值内的数据传回计算单元，异常数据发送至异常报警单元内，进行故障报警。Result display unit: used to judge the threshold value, send the data within the normal threshold value back to the calculation unit, and send the abnormal data to the abnormal alarm unit for fault alarm.
一种基于多维数据空间的轨道列车轴承故障诊断方法，包括以下步骤：A fault diagnosis method for rail train bearings based on multi-dimensional data space, comprising the following steps:
1)采集列车正常运行的样本数据；1) Collect the sample data of the normal operation of the train;
2)对采集到样本数据进行预处理和自动标签，形成参数训练样本集，构建分类模型并使用参数训练样本集进行训练，得到分类识别结果，即列车运行状态标签，并在参数训练样本集中划分运行状态对应的数据类；2) Preprocess and automatically label the collected sample data to form a parameter training sample set, build a classification model and use the parameter training sample set for training, and obtain the classification and recognition result, that is, the train running state label, and divide it into the parameter training sample set The data class corresponding to the running state;
3)对具有相同运行状态的数据类通过映射构建样本多维数据空间，并确定多维数据空间中每类正常运行状态的数据类对应的中心或质心；3) Build a sample multi-dimensional data space by mapping the data classes with the same operating state, and determine the center or centroid corresponding to each data class in the normal operating state in the multi-dimensional data space;
4)采集待检测的样本数据，并且采用训练好的分类模型识别得到分类识别结果后在待检测的样本数据中划分运行状态对应的数据类，并将其映射到样本多维数据空间中得到对应的检测样本；4) Collect the sample data to be detected, and use the trained classification model to identify and obtain the classification and recognition results, and then divide the data class corresponding to the running state in the sample data to be detected, and map it to the sample multi-dimensional data space to obtain the corresponding data. test samples;
5)计算检测样本数据点与中心或质心的距离，并与设定距离阈值相比较，根据比较结果进行故障报警并输出。5) Calculate the distance between the detected sample data point and the center or centroid, and compare it with the set distance threshold, and alarm and output the fault according to the comparison result.
所述的步骤1)中，样本数据包括轴承检测信号和列车状态识别辅助信息。In the step 1), the sample data includes the bearing detection signal and the auxiliary information for identifying the train state.
所述的轴承检测信号包括电机轴承振动信号、齿轮轴承振动信号和/或齿轮啮合振动信号。The bearing detection signal includes a motor bearing vibration signal, a gear bearing vibration signal and/or a gear meshing vibration signal.
所述的列车状态识别辅助信息包括列车载重、列车运行状态、列车姿态、列车速度和列车加速度信息，所述的列车运行状态包括启动、匀速、加速、减速、弯道和制动状态。The train state identification auxiliary information includes train weight, train running state, train attitude, train speed and train acceleration information, and the train running state includes start, constant speed, acceleration, deceleration, curve and braking state.
所述的步骤2)中，预处理具体包括以下步骤：In the described step 2), the preprocessing specifically includes the following steps:
21)数据整理：整理故障诊断样本集，剔除冗余和无效的数据；21) Data sorting: sorting out the fault diagnosis sample set, eliminating redundant and invalid data;
22)归一化处理：采用离差标准化进行归一化处理，以减少数据误差，保证后续分类识别结果的可靠性。22) Normalization processing: Normalization processing is performed using dispersion standardization to reduce data errors and ensure the reliability of subsequent classification and identification results.
所述的步骤2)中，分类模型采用SVM分类器或神经网络分类器。In the step 2), the classification model adopts SVM classifier or neural network classifier.
所述的步骤3)中，当具有相同运行状态的数据类为一维数据时，则以一维数据的中心作为故障判断依据，若具有相同运行状态的数据类为多维数据时，则以多维数据的质心作为故障判断依据。In the described step 3), when the data class with the same operating state is one-dimensional data, the center of the one-dimensional data is used as the basis for fault judgment, and if the data class with the same operating state is multi-dimensional data, the multi-dimensional data is used. The centroid of the data is used as the basis for fault judgment.
所述的一维数据的中心具体为一维数据的平均值，在一维数据空间中，以检测样本映射的点与中心之间的距离与距离阈值相比较，在距离阈值内的数据判定为正常数据，在距离阈值外的数据则判定为故障数据，并进行故障报警。The center of the one-dimensional data is specifically the average value of the one-dimensional data. In the one-dimensional data space, the distance between the point and the center of the detection sample mapping is compared with the distance threshold, and the data within the distance threshold is determined as: Normal data, data outside the distance threshold is judged as fault data, and a fault alarm is performed.
所述的多维数据由多个一维数据加权构成，在多维数据空间中，以检测样本映射的点与质心之间的欧式距离与距离阈值相比较，在距离阈值内的数据判定为正常数据，在距离阈值外的数据则判定为故障数据，并进行故障报警。The multi-dimensional data is composed of a plurality of one-dimensional data weights. In the multi-dimensional data space, the Euclidean distance between the point of the detection sample mapping and the centroid is compared with the distance threshold, and the data within the distance threshold is determined as normal data, The data outside the distance threshold is judged as fault data, and a fault alarm is performed.
与现有技术相比，本发明具有以下优点：Compared with the prior art, the present invention has the following advantages:
一、提高数据利用率：采用基于多维数据空间的诊断方法，可以在保障准确率的前提下，最大限度的利用更多维度的数据，降低故障误报率。1. Improve data utilization: Using the diagnosis method based on multi-dimensional data space can maximize the use of data in more dimensions and reduce the false alarm rate of faults on the premise of ensuring accuracy.
二、故障诊断可靠性高：在实际工况较为复杂，检测数据质量不高的情况下，通过对各检测信号对应权重构建多维数据空间加权矩阵，对具有相同运动状态的数据类创建样本多维坐标空间，并且，应用加权矩阵和自动标签分类，既考虑了整个数据集的特性，又消除了各数据集属性之间量纲的差异，并且在一定程度上减弱了噪声数据对距离度量的影响，将距离度量应用到入侵检测中，对复杂的数据集，可以排除次要数据类对结果造成的干扰，提高了系统的可靠性。2. High reliability of fault diagnosis: when the actual working conditions are complex and the quality of the detected data is not high, a multi-dimensional data space weighting matrix is constructed for the corresponding weights of each detection signal, and the multi-dimensional coordinates of the samples are created for the data classes with the same motion state. space, and the application of weighted matrix and automatic label classification not only considers the characteristics of the entire dataset, but also eliminates the dimensional difference between the attributes of each dataset, and to a certain extent reduces the influence of noise data on the distance metric, The application of distance metric to intrusion detection can eliminate the interference caused by secondary data types to the results for complex data sets, and improve the reliability of the system.
三、提高分类算法的计算速度：对数据集进行规范化处理，有助于加快分类算法的计算速度，并且可以帮助防止具有较大初始值域的属性与具有较小初始值域的属性相比权重过大，进而影响距离度量的准确性。3. Improve the calculation speed of the classification algorithm: Normalizing the data set helps to speed up the calculation speed of the classification algorithm, and can help prevent the attributes with a larger initial value range from being weighted compared to those with a smaller initial value range If it is too large, it will affect the accuracy of the distance measurement.
四、多层控制：本发明将基于多维数据空间的轨道列车轴承故障诊断系统分为多个层次的处理单元，然后根据列车实时运动状态，实时性强，可以根据列车运动状态不同造成的信号差异，进一步优化系统。4. Multi-layer control: The present invention divides the rail train bearing fault diagnosis system based on multi-dimensional data space into multiple levels of processing units, and then according to the real-time motion state of the train, the real-time performance is strong, and the signal difference caused by the different motion states of the train can be used. , to further optimize the system.
附图说明Description of drawings
图1为本发明轨道列车轴承故障诊断系统的结构框图。FIG. 1 is a structural block diagram of a bearing fault diagnosis system for rail trains according to the present invention.
图2为本发明轨道列车轴承故障诊断方法的实现流程图。FIG. 2 is a flow chart of the implementation of the fault diagnosis method for a bearing of a rail train according to the present invention.
具体实施方式Detailed ways
下面结合附图和具体实施例对本发明进行详细说明。The present invention will be described in detail below with reference to the accompanying drawings and specific embodiments.
实施例Example
本发明针对背景技术中提到的问题，提出了一种基于多维数据空间的轨道列车轴承故障诊断系统及方法，通过创建样本多维数据空间对轨道列车轴承部件的异常状态信息进行报警和定位，扩大了信息采集和应用的范围，提高了检测准确率和可靠性，具有较强的实用价值。In view of the problems mentioned in the background technology, the present invention proposes a fault diagnosis system and method for rail train bearings based on multi-dimensional data space. The scope of information collection and application is improved, the detection accuracy and reliability are improved, and it has strong practical value.
如图1所示，本发明包括一种轨道列车轴承故障诊断系统框图，包括数据采集单元、数据发送单元、运行状态识别单元、计算单元、结果显示单元、异常报警单元。数据发送单元输入端为数据采集单元，其输出端为运行状态识别单元，计算单元包含质心计算模块、中心值计算模块、距离计算模块、差值计算模块，输入端为运行状态识别单元，将振动检测数据分别输出到质心计算模块和中心值计算模块中，结果显示单元的输入端为计算单元，诊断结果正常则将数据集反馈给计算单元重新计算，将异常结果输出至异常报警单元。As shown in Figure 1, the present invention includes a block diagram of a rail train bearing fault diagnosis system, including a data acquisition unit, a data transmission unit, a running state identification unit, a calculation unit, a result display unit, and an abnormal alarm unit. The input end of the data sending unit is a data acquisition unit, and its output end is a running state identification unit. The calculation unit includes a centroid calculation module, a center value calculation module, a distance calculation module, and a difference value calculation module. The detection data is output to the centroid calculation module and the central value calculation module respectively. The input end of the result display unit is the calculation unit. If the diagnosis result is normal, the data set is fed back to the calculation unit for recalculation, and the abnormal result is output to the abnormal alarm unit.
数据采集单元用以采集轴承检测信号和列车状态识别辅助信息数据；数据发送单元用以将采集到的数据发送至运行状态识别单元；运行状态识别单元用以接收数据发送单元传送的数据，并对采集到的数据进行处理，获得列车运行状态；计算单元用以计算质心和中心值，并分别计算各数据点与质心和中心值之间的距离和差值，将结果传送至结果显示单元；结果显示单元用以进行阈值判断，将在正常阈值内的数据传回计算单元，异常信息传送至异常报警单元内，进行故障报警。The data acquisition unit is used to collect the bearing detection signal and the auxiliary information data of the train state identification; the data transmission unit is used to send the collected data to the operation state identification unit; the operation state identification unit is used to receive the data transmitted by the data transmission unit, and to The collected data is processed to obtain the train running state; the calculation unit is used to calculate the centroid and center value, and calculate the distance and difference between each data point and the centroid and center value respectively, and transmit the result to the result display unit; The display unit is used for judging the threshold value, sending the data within the normal threshold value back to the computing unit, and sending the abnormal information to the abnormal alarm unit for fault alarming.
图2为本发明轨道列车轴承故障诊断方法的实现流程图，该方法基于数据空间判定方法建立轨道列车轴承的故障诊断模型，针对复杂工况下轴承状态采集信息使用率不足，故障判断可靠性不足的情况，引入多维检测信号坐标空间的阈值判断方法进行精确分类，包括以下步骤：Fig. 2 is a flow chart of the implementation of the fault diagnosis method for rail train bearings according to the present invention. The method establishes a fault diagnosis model for rail train bearings based on the data space determination method, and the utilization rate of the information collected for the bearing state under complex working conditions is insufficient, and the reliability of fault judgment is insufficient. In the case of , the threshold judgment method of multi-dimensional detection signal coordinate space is introduced for accurate classification, including the following steps:
步骤1、采集样本，包括轴承检测信号和列车状态识别辅助信息；Step 1. Collect samples, including bearing detection signals and auxiliary information for train status identification;
步骤2、对原始数据进行预处理和自动标签；Step 2. Preprocess and automatically label the original data;
步骤3、分类识别算法对获得的数据进行训练，识别列车运行状态，确定数据类；Step 3. The classification and identification algorithm trains the obtained data, identifies the train running state, and determines the data class;
步骤4、基于已进行列车运行状态识别获得的具有相同标签的数据类，依据各检测信号对应权重构建多维数据空间加权矩阵并获得同步检测数据向量空间坐标，创建样本多维数据空间，包括一维检测信号向量和多维检测信号向量组；Step 4. Based on the data classes with the same label obtained by the recognition of the train running state, construct a multi-dimensional data space weighting matrix according to the corresponding weights of each detection signal, obtain the synchronous detection data vector space coordinates, and create a sample multi-dimensional data space, including one-dimensional detection. Signal vector and multi-dimensional detection signal vector group;
步骤5、对于一维检测信号向量，计算一维向量的中心值，以及检测样本各数据点与中心值的差值，进行待检信号与单参数中心值差值阈值的比较判断；对于多维检测信号向量组，获得样本多维数据空间计算质心坐标，并计算待检单次同步采集信号各数据点与质心的欧式距离，并将其与质心距离阈值进行比较判断；Step 5. For the one-dimensional detection signal vector, calculate the central value of the one-dimensional vector, and the difference between each data point of the detection sample and the central value, and compare and judge the difference threshold between the signal to be detected and the central value of the single parameter; for multi-dimensional detection Signal vector group, obtain the sample multi-dimensional data space to calculate the centroid coordinates, and calculate the Euclidean distance between each data point and the centroid of the single synchronous acquisition signal to be tested, and compare it with the centroid distance threshold;
步骤6、获得异常检测结果，报警并输出异常检测点。Step 6. Obtain abnormal detection results, alarm and output abnormal detection points.
各步骤的详细介绍如下：The details of each step are as follows:
步骤1、采集样本：通过传感器实时采集列车运行过程中的轴承检测信号，主要包括单个或多个电机轴承振动信号、单个或多个齿轮轴承振动信号、单个或多个齿轮啮合振动信号(对应一维或多维检测信号)，列车状态识别辅助信息，主要包括列车载重、列车运行状态记录、列车姿态、列车速度、列车加速度。Step 1. Collect samples: real-time acquisition of bearing detection signals during train operation through sensors, mainly including single or multiple motor bearing vibration signals, single or multiple gear bearing vibration signals, single or multiple gear meshing vibration signals (corresponding to a Dimension or multi-dimensional detection signal), auxiliary information for train status identification, mainly including train load, train running status record, train attitude, train speed, and train acceleration.
步骤2、列车状态识别：采用分类识别算法对列车运行状态进行识别，对步骤1中获得的原始采集样本进行预处理，数据预处理主要包括：Step 2. Train state identification: use the classification and identification algorithm to identify the train running state, and preprocess the original collection samples obtained in step 1. The data preprocessing mainly includes:
步骤2.1、数据整理：整理故障诊断样本集，剔除冗余、无效的数据；Step 2.1, data sorting: sorting out the fault diagnosis sample set, eliminating redundant and invalid data;
步骤2.2、归一化处理：由于不同检测信号之间的量纲以及数量级都有不同，因此需要对数据进行归一化处理，减少数据分析结果的偏差，保证后续结果的可靠性，本发明对输入变量采用的归一化方法是离差标准化，将原始数据归一化到[0，1]之间，针对归一化前的数据X(i)，经过归一化处理后数据X(i)′，Step 2.2, normalization processing: Since the dimension and order of magnitude between different detection signals are different, it is necessary to normalize the data to reduce the deviation of data analysis results and ensure the reliability of subsequent results. The normalization method used for the input variable is dispersion normalization, which normalizes the original data to between [0, 1]. For the data X(i) before normalization, the data X(i) after normalization is processed. )',
计算公式如下：Calculated as follows:
其中i＝1,2,...,i，为样本中单一采集信号的采集次数。where i=1,2,...,i, is the acquisition times of a single acquisition signal in the sample.
步骤3、选取识别算法分别对不同运动状态下的参数训练样本进行训练，得到对采集数据进行列车运行状态识别划分的分类模型，对待检测的样本数据采用获得的分类模型进行列车运行状态识别，划分具体数据类。Step 3: Select the recognition algorithm to train the parameter training samples in different motion states respectively, and obtain a classification model for identifying and dividing the train running state on the collected data, and use the obtained classification model to identify the train running state for the sample data to be detected. specific data class.
步骤4、创建样本多维数据空间：采用步骤2中已进行列车运行状态识别的具有相同标签的数据类，依据各检测信号对应权重，构建多维数据加权矩阵并获得同步检测数据向量在多维数据空间中的对应坐标，对具有相同运动状态的数据类创建样本多维数据空间。Step 4. Create a sample multi-dimensional data space: Use the data class with the same label that has been identified in step 2 for the train running state, construct a multi-dimensional data weighting matrix according to the corresponding weight of each detection signal, and obtain the synchronous detection data vector in the multi-dimensional data space. The corresponding coordinates of , create a sample multi-dimensional data space for data classes with the same motion state.
具体如下：一维检测信号构成的向量，The details are as follows: the vector formed by the one-dimensional detection signal,
[X(t1),X(t2),…,X(tn)][X(t1 ),X(t2 ),…,X(tn )]
其中，[t1,t2,…,tn]代表同步采集时间序列。由多组一维检测信号向量构成如下多维数据加权矩阵，Among them, [t1 ,t2 ,…,tn ] represents the synchronous acquisition time series. The following multi-dimensional data weighting matrix is composed of multiple sets of one-dimensional detection signal vectors,
其中，1,2,3,…,j为依据样本创建的多维数据空间维度，[ω1,ω2,…,ωn]是各维检测参数的权重。Among them, 1, 2, 3,...,j are the multi-dimensional data space dimensions created according to the sample, and [ω1 ,ω2 ,...,ωn ] is the weight of each dimension detection parameter.
步骤5、基于多维数据空间判断。Step 5: Judging based on the multidimensional data space.
步骤5.1、计算一维数据中心值：Step 5.1, calculate the one-dimensional data center value:
令X(t1),X(t2),…,X(tn)为参数训练样本的j维检测信号的n个观测值，一维数据的中心值采用均值或中值作为数值度量，依据均值计算公式
步骤5.2、依据参数训练样本集计算多维数据空间质心坐标，计算公式为：Step 5.2. Calculate the coordinates of the centroid of the multi-dimensional data space according to the parameter training sample set. The calculation formula is:
其中，k∈1,2,…,n。则由参数训练样本集获得的质心坐标
待检的数据映射到多维空间中的坐标为(ω1X1，ω2X2，…，ωjXj)，采用欧式距离计算待检测的样本数据与质心的距离值，即：The coordinates of the data to be tested mapped to the multi-dimensional space are (ω1 X1 , ω2 X2 , ..., ωj Xj ), and the Euclidean distance is used to calculate the distance between the sample data to be tested and the centroid, namely:
步骤5.3、参数阈值判断：包括一维阈值Ssj判断，和多维阈值Sm判断。Step 5.3, parameter threshold judgment: including one-dimensional threshold value Ssj judgment, and multi-dimensional threshold value Sm judgment.
基于参数训练样本集，及列车正常运行状态下检测到的轴承检测振动信号，在单一维度上，设置对应的正常阈值范围
基于参数训练样本集，在多维数据空间中设置对应的正常阈值，即以质心为中心，距离质心Sm范围内。如果待检测的样本数据与质心的欧式距离大于Sm，判断为异常情况。Based on the parameter training sample set, the corresponding normal threshold is set in the multi-dimensional data space, that is, the center of mass is taken as the center, and the distance from the center of mass is within the range of Sm. If the Euclidean distance between the sample data to be detected and the centroid is greater than Sm, it is judged as an abnormal situation.
步骤6、获得异常检测结果，报警并输出异常检测点。Step 6. Obtain abnormal detection results, alarm and output abnormal detection points.
以上所述仅为本发明的较佳实例，并不用以限制本发明，凡在本发明的精神和原The above descriptions are only preferred examples of the present invention, and are not intended to limit the present invention.
则内所作的任何修改、等同替换和改进等，均应包含在本发明的保护范围之内。Any modifications, equivalent replacements and improvements made in the above shall be included within the protection scope of the present invention.
Claims (3)
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010384367.3ACN111539374B (en) | 2020-05-07 | 2020-05-07 | Fault diagnosis method of rail train bearing based on multi-dimensional data space |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010384367.3ACN111539374B (en) | 2020-05-07 | 2020-05-07 | Fault diagnosis method of rail train bearing based on multi-dimensional data space |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111539374A CN111539374A (en) | 2020-08-14 |
| CN111539374Btrue CN111539374B (en) | 2022-05-20 |
Family
ID=71979182
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010384367.3AActiveCN111539374B (en) | 2020-05-07 | 2020-05-07 | Fault diagnosis method of rail train bearing based on multi-dimensional data space |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111539374B (en) |
Families Citing this family (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113156913B (en)* | 2021-02-05 | 2024-08-06 | 深圳大雷汽车检测股份有限公司 | ABS fault diagnosis system and method |
| CN112836896A (en)* | 2021-03-03 | 2021-05-25 | 西门子工厂自动化工程有限公司 | Method for maintaining equipment and system for maintaining equipment |
| CN113569475B (en)* | 2021-07-21 | 2023-09-01 | 上海工程技术大学 | A Fault Diagnosis System for Subway Axlebox Bearings Based on Digital Twin Technology |
| CN115809435B (en)* | 2023-02-06 | 2023-05-12 | 山东星科智能科技股份有限公司 | Automobile operation fault identification method based on simulator |
| CN117332233B (en)* | 2023-10-07 | 2024-05-31 | 江苏丰昌机电科技有限公司 | A motor intelligent maintenance system |
| CN117943891B (en)* | 2024-03-22 | 2024-09-06 | 济南二机床集团有限公司 | Method, device, equipment and medium for detecting fault of electric spindle bearing of numerical control machine tool |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2018015127A1 (en)* | 2016-07-20 | 2018-01-25 | Thales Austria Gmbh | System and method for ascertaining at least one state of a train |
| CN110111446A (en)* | 2019-01-25 | 2019-08-09 | 东北大学 | A kind of adaptive acquisition system of high-speed rail train data and method based on operating condition |
Family Cites Families (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| BR9814278A (en)* | 1997-09-12 | 2005-08-16 | New York Air Brake Corp | Train operation and training optimization method |
| AR091412A1 (en)* | 2013-06-11 | 2015-02-04 | Ypf Sa | DEVICE AND METHOD FOR DIAGNOSING ANOMAL SITUATIONS IN PROCESSES |
| CN104392071B (en)* | 2014-12-12 | 2017-09-29 | 北京交通大学 | A kind of bullet train system security assessment method based on complex network |
| CN104931263B (en)* | 2015-06-18 | 2016-10-12 | 东南大学 | A kind of Method for Bearing Fault Diagnosis based on symbolization probabilistic finite state machine |
| CN109100143A (en)* | 2018-07-06 | 2018-12-28 | 华中科技大学 | Fault Diagnosis of Roller Bearings and equipment based on CEEMDAN and CFSFDP |
| CN109033582A (en)* | 2018-07-12 | 2018-12-18 | 西安英特迈思信息科技有限公司 | A kind of bullet train multi parameter intallingent threshold value criterion |
| CN110503004B (en)* | 2019-07-29 | 2022-03-22 | 七彩安科智慧科技有限公司 | On-line judging method for operating state of switching power supply |
- 2020
- 2020-05-07CNCN202010384367.3Apatent/CN111539374B/enactiveActive
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2018015127A1 (en)* | 2016-07-20 | 2018-01-25 | Thales Austria Gmbh | System and method for ascertaining at least one state of a train |
| CN110111446A (en)* | 2019-01-25 | 2019-08-09 | 东北大学 | A kind of adaptive acquisition system of high-speed rail train data and method based on operating condition |
Non-Patent Citations (1)
| Title |
|---|
| Triplet Loss Guided Adversarial Domain Adaptation for Bearing Fault Diagnosis;Xiaodong Wang et al;《sensors》;20200106;1-19页* |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111539374A (en) | 2020-08-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111539374B (en) | Fault diagnosis method of rail train bearing based on multi-dimensional data space | |
| Bai et al. | An optimized railway fastener detection method based on modified Faster R-CNN | |
| CN111259990B (en) | Seawater pump fault monitoring device and diagnosis method based on multi-source information fusion | |
| CN111562108A (en) | An Intelligent Fault Diagnosis Method of Rolling Bearing Based on CNN and FCMC | |
| CN110287552A (en) | Method and system for fault diagnosis of motor bearing based on improved random forest algorithm | |
| CN111461185A (en) | Driving behavior analysis method based on improved K-means | |
| CN111917766B (en) | Method for detecting communication abnormity of vehicle-mounted network | |
| CN114564987B (en) | Rotary machine fault diagnosis method and system based on graph data | |
| CN115356133B (en) | Bogie operation condition online monitoring, real-time classifying and identifying algorithm | |
| CN103471865A (en) | Train suspension system failure isolation method based on LDA method | |
| CN114861719A (en) | High-speed train bearing fault diagnosis method based on ensemble learning | |
| CN110728459A (en) | Travel mode identification system, method and device and model training method and device | |
| Zhou et al. | Reliable and intelligent fault diagnosis with evidential VGG neural networks | |
| CN113758709A (en) | Rolling bearing fault diagnosis method and system combining edge computing and deep learning | |
| CN117668689A (en) | Rail transit train bearing fault diagnosis method and device based on meta-learning algorithm | |
| CN114492642A (en) | Mechanical fault online diagnosis method for multi-scale element depth residual shrinkage network | |
| Duan et al. | Rolling bearing fault diagnosis method based on multi-information fusion characteristics under complex working conditions | |
| CN118762418B (en) | A method and system for monitoring the running status of underground railcar | |
| CN111351655A (en) | Method for diagnosing gear fault in locomotive gear box | |
| CN119441885A (en) | A method for constructing a generalized fault diagnosis model in a single-source domain under time-varying conditions and its application | |
| Yu et al. | Research on the fault diagnosis technology of the railway vehicle transmission system based on a convolutional neural network | |
| CN116485284A (en) | Rail transit comprehensive joint debugging evaluation system and method | |
| CN116580292A (en) | Track structure state detection method, computer equipment and medium | |
| CN119961872B (en) | Vehicle condition monitoring method, device, terminal equipment and medium based on automobile | |
| CN118194178B (en) | Intelligent analysis method for operation data of ship energy management system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |