技术领域technical field
本发明涉及故障诊断领域，特别是涉及一种电力驱动交通工具的故障诊断方法和系统。The invention relates to the field of fault diagnosis, in particular to a fault diagnosis method and system for an electric drive vehicle.
背景技术Background technique
在电力驱动交通工具的单体电压统计过程中，可能发生单体电压故障，导致个别单体电压异常，从而导致对中心值的统计计算中出现偏差。因此在电力交通工具进行电池数据处理时，会因个别单体电压异常而导致得到的数据中可能会有一些远离中心值的故障数据。当我们希望得到一组数据的中心值时，这些故障数据总会干扰真实中心值的计算，从而影响故障分析。During the statistical process of cell voltage of electric drive vehicles, cell voltage faults may occur, resulting in abnormal voltage of individual cells, resulting in deviations in the statistical calculation of the central value. Therefore, when an electric vehicle performs battery data processing, there may be some fault data far away from the central value in the obtained data due to the abnormal voltage of individual cells. When we want to get the central value of a set of data, these fault data will always interfere with the calculation of the real central value, thereby affecting the fault analysis.
另外，传统的故障分析大多是先发现故障，发现故障后寻找导致故障的原因，之后找出这一个或多个原因和这个故障之间的确定关系，进而研究故障的发生概率等，但参数间的耦合复杂，建模难等问题直接影响数据的处理时效性。尽管本领域研究故障诊断方法众多，但如何运用统计学方法分析故障发生原因，在目前现有技术中仍为空白。In addition, most of the traditional fault analysis is to find the fault first, find the cause of the fault after finding the fault, and then find out the definite relationship between the one or more reasons and the fault, and then study the probability of the fault, etc., but the parameters between Problems such as complex coupling and difficult modeling directly affect the timeliness of data processing. Although there are many fault diagnosis methods in this field, how to use statistical methods to analyze the cause of faults is still blank in the current prior art.
发明内容Contents of the invention
本发明的目的是提供一种电力驱动交通工具的故障诊断方法和系统，建立3σ多层次数据筛选模型，运用高斯分布的概率特性，对无故障数据进行集中筛选，对故障数据高效、精确的定位并剔除，已解决传统由于参数间的耦合复杂、建模难、影响数据的处理时效性的问题。The purpose of the present invention is to provide a fault diagnosis method and system for electric-driven vehicles, establish a 3σ multi-level data screening model, use the probability characteristics of Gaussian distribution, conduct centralized screening of non-fault data, and efficiently and accurately locate fault data And eliminated, it has solved the traditional problems of complex coupling between parameters, difficult modeling, and affecting the timeliness of data processing.
为实现上述目的，本发明提供了如下方案：To achieve the above object, the present invention provides the following scheme:
一种电力驱动交通工具的故障诊断方法，包括步骤：A fault diagnosis method for an electric drive vehicle, comprising the steps of:
步骤A、获取被监控交通工具的电池的运行参数数据，对所述运行参数数据进行筛选整理得到被监控交通工具的电池单体的电压的第一数据矩阵；Step A. Obtain the operating parameter data of the battery of the monitored vehicle, and filter and arrange the operating parameter data to obtain the first data matrix of the voltage of the battery cell of the monitored vehicle;
步骤B、对所述第一数据矩阵采用3σ多层次筛选算法得到每个电池单体的故障数；Step B, using a 3σ multi-level screening algorithm on the first data matrix to obtain the number of faults of each battery cell;
步骤C、根据所述故障数计算每个电池单体的故障概率；Step C, calculating the failure probability of each battery cell according to the number of failures;
步骤D、判断每个电池单体的故障概率是否超过第二给定值，是，则确定故障发生。Step D, judging whether the failure probability of each battery cell exceeds a second given value, if yes, determine that a failure has occurred.
其中，所述的数据矩阵为m行n列矩阵，步骤B包括步骤：Wherein, the data matrix is a matrix of m rows and n columns, and step B includes the steps of:
B1、判断m行数据矩阵是否已完成循环，是，则执行步骤B8，否，则执行步骤B2；B1, judging whether the m row data matrix has completed the cycle, if yes, execute step B8, otherwise, execute step B2;
B2、计算第一数据矩阵的行向量的第一均值μ1和第一标准差σ1，建立第一高斯分布；B2. Calculate the first mean value μ1 and the first standard deviation σ1 of the row vector of the first data matrix, and establish the first Gaussian distribution;
B3、设置第一故障矩阵R1记录所述第一数据矩阵中超过3σ1的元素的位置；B3, the first failure matrix R1 is set to record the positions of elements exceeding 3σ1 in the first data matrix;
B4、剔除所述第一数据矩阵中第一故障矩阵R1记录的对应位置的元素获得第二数据矩阵，再重新计算第二数据矩阵的行向量的第二均值μ2和第二标准差σ2；B4, get rid of the element of the corresponding position recorded in the first fault matrix R1 in the first data matrix to obtain the second data matrix, and then recalculate the second mean value μ2 and the second standard deviation σ2 of the row vector of the second data matrix ;
B5、设置第二故障矩阵R2记录所述第一数据矩阵中超过3σ2的元素的位置；B5, the second failure matrix R2 is set to record the positions of elements exceeding3σ2 in the first data matrix;
B6、判断σ1-σ2的差值是否大于第一给定值，是，则μ1＝μ2，σ1＝σ2，R1＝R2，并返回到步骤B4；否，则执行步骤B7；B6. Judging whether the difference between σ1 -σ2 is greater than the first given value, if yes, then μ1 = μ2 , σ1 =σ2 , R1 = R2, and return to step B4; otherwise, execute step B7 ;
B7、输出第二数据矩阵的行向量的第二均值μ2和第二标准差σ2，建立第二高斯分布；B7. Output the second mean value μ2 and the second standard deviation σ2 of the row vector of the second data matrix, and establish the second Gaussian distribution;
B8、将第二故障矩阵R2中的每一列向量相加得到每个电池单体的故障数。B8. Adding up each column vector in the second failure matrix R2 to obtain the number of failures of each battery cell.
其中，每个电池单体的故障概率为：m为所述数据矩阵的行数；Among them, the failure probability of each battery cell is: m is the number of rows of the data matrix;
其中，步骤D包括步骤：Wherein, step D comprises steps:
D1、根据所述故障概率判断是否超过第二给定值，是，则执行步骤D2，否，则执行步骤D3；D1. Judging whether the failure probability exceeds the second given value, if yes, execute step D2, and if no, execute step D3;
D2、对电池单体进行报错；D2. Report an error to the battery cell;
D3、对超过第二给定值的电池单体统计规律，并得到相应的结果。D3. Statistical rules for battery cells exceeding the second given value, and obtaining corresponding results.
本发明还提供了一种电力驱动交通工具的故障诊断系统，包括获取模块、3σ筛选模块、求取故障概率模块、故障判断模块；The present invention also provides a fault diagnosis system for electric-driven vehicles, including an acquisition module, a 3σ screening module, a fault probability calculation module, and a fault judgment module;
所述获取模块，用于获取被监控交通工具的电池的运行参数数据，对所述运行参数数据进行筛选整理得到被监控交通工具的电池单体的电压的第一数据矩阵；The acquiring module is used to acquire the operating parameter data of the battery of the monitored vehicle, and filter and arrange the operating parameter data to obtain the first data matrix of the voltage of the battery cell of the monitored vehicle;
所述3σ筛选模块，用于对所述第一数据矩阵采用3σ多层次筛选算法得到每个电池单体的故障数；The 3σ screening module is configured to use a 3σ multi-level screening algorithm for the first data matrix to obtain the number of faults of each battery cell;
所述求取故障概率模块，用于根据所述故障数计算每个电池单体的故障概率；The module for calculating the failure probability is used to calculate the failure probability of each battery cell according to the number of failures;
所述故障判断模块，用于判断每个电池单体的故障概率是否超过第二给定值，是，则确定故障发生。The failure judging module is used to judge whether the failure probability of each battery cell exceeds a second given value, and if yes, determine that a failure has occurred.
所述3σ筛选模块，具体用于：判断m行数据矩阵是否已完成循环，否，则计算第一数据矩阵的行向量的第一均值μ1和第一标准差σ1，建立第一高斯分布，记录所述第一数据矩阵中超过3σ1的元素的位置，剔除所述第一数据矩阵中第一故障矩阵R1记录的对应位置的元素获得第二数据矩阵，再重新计算第二数据矩阵的行向量的第二均值μ2和第二标准差σ2，记录所述第一数据矩阵中超过3σ2的元素的位置，直至σ1-σ2的差值是小于第一给定值为止，输出第二数据矩阵的行向量的第二均值μ2和第二标准差σ2，建立第二高斯分布，当m行数据矩阵已完成循环时，将第二故障矩阵R2中的每一列向量相加得到每个电池单体的故障数。The 3σ screening module is specifically used for: judging whether the m-row data matrix has completed the cycle, otherwise, calculating the first mean value μ1 and the first standard deviation σ1 of the row vector of the first data matrix, and establishing the first Gaussian distribution , record the positions of the elements exceeding 3σ1 in thefirst data matrix, remove the elements of the corresponding positions recorded by the first failure matrix R1 in the first data matrix to obtain the second data matrix, and recalculate the second data matrix The second mean value μ2 and the second standard deviation σ2 of the row vector, record the positions of elements exceeding 3σ2 in the first data matrix, until the difference between σ1 -σ2 is less than the first given value, output The second mean value μ2 and the second standard deviation σ2 of the row vectors of the second data matrix, establish the second Gaussian distribution, when the m-row data matrix has completed the cycle, add each column vector in the second fault matrix R2 Get the number of failures per battery cell.
所述求取故障概率模块，用于采用如下式子计算每个电池单体的故障概率：m为所述数据矩阵的行数。The module for obtaining the failure probability is used to calculate the failure probability of each battery cell using the following formula: m is the number of rows of the data matrix.
所述故障判断模块包括第二逻辑判断单元、故障报错单元、统计规律单元；The fault judgment module includes a second logical judgment unit, a fault error reporting unit, and a statistical law unit;
所述第二逻辑判断单元，用于根据所述故障概率判断是否超过第二给定值；The second logic judging unit is configured to judge whether the failure probability exceeds a second given value according to the failure probability;
所述故障报错单元，用于对电池单体进行报错；The fault error reporting unit is used to report an error to the battery cell;
所述统计规律单元，用于对超过第二给定值的电池单体统计规律，并得到相应的结果。The statistical law unit is used for statistical law of the battery cells exceeding the second given value, and obtain corresponding results.
根据本发明提供的具体实施例，本发明公开了以下技术效果：According to the specific embodiments provided by the invention, the invention discloses the following technical effects:
本发明提供了一种电力驱动交通工具的故障诊断方法和系统，首先，通过获取电力驱动装置监控的电池数据进行筛选整理得到数据矩阵，保证了数据的真实性、实时性、可靠性；其次，采用高斯分布的概率特性与3σ置信区间相结合而构成的3σ多层次筛选对无故障数据进行筛选，对于一些低阶故障数据可实现一次筛选就能去除掉所有超过3σ置信区间的数据，对于一些多阶故障数据或存在数量级差别较大的故障数据可多次筛选和剔除，获得最接近真实数据的中心值，处理效率高；最后，通过调整第二给定值，可实现对故障值的筛选，进而做到精确诊断和控制，为后期车辆数目、车辆种类、不同时间等多维度的大数据处理，提供了高效的算法。The present invention provides a fault diagnosis method and system for an electric drive vehicle. First, the data matrix is obtained by screening and sorting the battery data monitored by the electric drive device, which ensures the authenticity, real-time and reliability of the data; secondly, The 3σ multi-level screening composed of the combination of the probability characteristics of Gaussian distribution and the 3σ confidence interval is used to screen the non-fault data. For some low-order fault data, all the data exceeding the 3σ confidence interval can be removed by one screening. For some Multi-level fault data or fault data with large magnitude differences can be screened and eliminated multiple times to obtain the central value closest to the real data, with high processing efficiency; finally, by adjusting the second given value, the screening of fault values can be realized , and then achieve accurate diagnosis and control, and provide an efficient algorithm for multi-dimensional big data processing such as the number of vehicles, vehicle types, and different times in the later stage.
附图说明Description of drawings
为了更清楚地说明本发明实施例或现有技术中的技术方案，下面将对实施例中所需要使用的附图作简单地介绍，显而易见地，下面描述中的附图仅仅是本发明的一些实施例，对于本领域普通技术人员来讲，在不付出创造性劳动性的前提下，还可以根据这些附图获得其他的附图。In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the accompanying drawings required in the embodiments. Obviously, the accompanying drawings in the following description are only some of the present invention. Embodiments, for those of ordinary skill in the art, other drawings can also be obtained according to these drawings without paying creative labor.
图1为本发明实施例电力驱动交通工具的故障诊断方法的主体流程图；Fig. 1 is a main flow chart of a fault diagnosis method for an electric drive vehicle according to an embodiment of the present invention;
图2为本发明实施例电力驱动交通工具的故障诊断方法的具体流程图；2 is a specific flow chart of a fault diagnosis method for an electric drive vehicle according to an embodiment of the present invention;
图3为本发明实施例电力驱动交通工具的故障诊断系统结构图；3 is a structural diagram of a fault diagnosis system of an electric drive vehicle according to an embodiment of the present invention;
图4为电池的单体电压曲线图；Fig. 4 is the monomer voltage graph of battery;
图5为第一高斯分布曲线图；Fig. 5 is the first Gaussian distribution curve figure;
图6为第二高斯分布曲线图；Fig. 6 is the second Gaussian distribution curve figure;
图7为两次高斯分布比较曲线图。Fig. 7 is a comparison curve of two Gaussian distributions.
具体实施方式Detailed ways
下面将结合本发明实施例中的附图，对本发明实施例中的技术方案进行清楚、完整地描述，显然，所描述的实施例仅仅是本发明一部分实施例，而不是全部的实施例。基于本发明中的实施例，本领域普通技术人员在没有做出创造性劳动前提下所获得的所有其他实施例，都属于本发明保护的范围。The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only some, not all, embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by persons of ordinary skill in the art without making creative efforts belong to the protection scope of the present invention.
本发明的目的是提供一种电力驱动交通工具的故障诊断方法和系统。The object of the present invention is to provide a fault diagnosis method and system for electric drive vehicles.
为使本发明的上述目的、特征和优点能够更加明显易懂，下面结合附图和具体实施方式对本发明作进一步详细的说明。In order to make the above objects, features and advantages of the present invention more comprehensible, the present invention will be further described in detail below in conjunction with the accompanying drawings and specific embodiments.
参见图1和图2所示，本发明实施例提供一种电力驱动交通工具的故障诊断方法。Referring to Fig. 1 and Fig. 2, an embodiment of the present invention provides a fault diagnosis method for an electric drive vehicle.
步骤S1，获取被监控交通工具的电池的运行参数数据，对所述运行参数数据进行筛选整理得到被监控交通工具的电池单体的电压的数据矩阵。In step S1, the operating parameter data of the battery of the monitored vehicle is obtained, and the operating parameter data is screened to obtain a data matrix of the voltage of the battery cells of the monitored vehicle.
首先，获取某一段时间内的电池的运行参数数据，这些数据主要来源于电力驱动交通工具监控中心，电池的运行参数主要包括电池系统总电压和总电流、SOC状态、电池单体电压、电池包特征点温度等。First, obtain the operating parameter data of the battery within a certain period of time. These data mainly come from the monitoring center of electric vehicles. The operating parameters of the battery mainly include the total voltage and total current of the battery system, SOC status, battery cell voltage, battery pack Feature point temperature, etc.
其次，对电力驱动交通监控平台进行车型和公里数的筛选，确保几个对比组的车辆种类，车辆公里数、放电时SOC变化范围基本相同，避免其他因素造成干扰，影响检测的准确性，还要确保汽车传感器正常工作。Secondly, screen the vehicle type and mileage of the electric drive traffic monitoring platform to ensure that the vehicle types, vehicle mileage, and SOC variation range during discharge are basically the same in several comparison groups, so as to avoid interference caused by other factors and affect the accuracy of detection. Make sure the car sensors are working properly.
最后，将电池的运行参数导入数值计算工具中进行整理，获得被监控交通工具的电池单体的电压的第一数据矩阵，第一数据矩阵为m行n列。Finally, the operating parameters of the battery are imported into a numerical calculation tool for sorting, and the first data matrix of the voltage of the battery cells of the monitored vehicle is obtained, and the first data matrix has m rows and n columns.
本申请中的n既表示第一数据矩阵的列数，还表示电池单体的数目；m既表示第一数据矩阵的行数，还表示某一时间段内进行m次电压采样。In this application, n not only represents the number of columns of the first data matrix, but also represents the number of battery cells; m not only represents the number of rows of the first data matrix, but also represents m times of voltage sampling within a certain period of time.
具体其数据矩阵可表示为：Specifically, its data matrix can be expressed as:
其中，Ut0,j-Utm,j表示同一电池单体不同时刻的电压值，Uti,0-Uti,n表示不同单体在ti时刻的电压，U0表示为第一数据矩阵，第i行相当于第ti时刻对n个电池单体进行电压采样，其中1≤i≤m,t1≤ti≤tm,1≤j≤n。 Among them, Ut0,j -Utm,j represents the voltage value of the same battery cell at different times, Uti,0 -Uti,n represents the voltage of different cells at time ti, U0 represents the first data matrix, The i-th row is equivalent to sampling the voltage of n battery cells at the ti-th time, where 1≤i≤m, t1≤ti≤tm, 1≤j≤n.
步骤S2、对所述第一数据矩阵采用3σ多层次筛选算法得到每个电池单体的故障数，其具体步骤如下：Step S2, using a 3σ multi-level screening algorithm on the first data matrix to obtain the number of faults of each battery cell, the specific steps are as follows:
步骤S201、判断m行数据矩阵是否已完成循环，是，则执行步骤S208，否，则执行步骤S202。Step S201, judging whether the m-row data matrix has completed the cycle, if yes, execute step S208, and if no, execute step S202.
本申请通过比较不同时刻单体电压采样值共同判断单体电压是否发生故障，所以需要判断该矩阵m行是否已经完成循环。In this application, by comparing the sampling values of the cell voltage at different times, it is judged whether the cell voltage is faulty, so it is necessary to judge whether the m rows of the matrix have completed the cycle.
步骤S202、计算第一数据矩阵的行向量第一均值μ1和第一标准差σ1，建立第一高斯分布。Step S202. Calculate the first mean value μ1 and the first standard deviation σ1 of the row vectors of the first data matrix, and establish the first Gaussian distribution.
高斯分布(Gaussian distribution)，又称正态分布(Normal distribution)，是一种非常普遍的连续概率函数。Gaussian distribution, also known as normal distribution, is a very common continuous probability function.
若随机变量X服从一个数学期望为μ，方差为σ2的概率分布时，其高斯分布函数为：If the random variable X obeys a probability distribution with a mathematical expectation of μ and a variance ofσ2 , its Gaussian distribution function is:
其图像为钟形曲线，其中μ决定了钟形曲线的中心位置，标准差σ决定了分布尺度，所以又称μ为位置参数，标准差σ为尺度参数。X服从正态分布记为：X～N(μ，σ2)。Its image is a bell curve, where μ determines the center position of the bell curve, and the standard deviation σ determines the distribution scale, so μ is also called the position parameter, and the standard deviation σ is the scale parameter. X obeys normal distribution and is recorded as: X～N(μ, σ2 ).
通过数值计算工具计算第一数据矩阵的第一均值μ1和第一标准差σ1，并将上述的第一均值μ1和第一标准差σ1带入高斯分布函数即可绘制第一高斯分布曲线图，其服从正态分布(μ1，σ12)。Calculate the first mean value μ1 and the first standard deviation σ1 of the first data matrix through numerical calculation tools, and bring the above-mentioned first mean value μ1 and first standard deviation σ1 into the Gaussian distribution function to draw the first Gaussian Distribution curve graph, which obeys normal distribution (μ1 , σ12 ).
步骤S203、设置第一故障矩阵R1记录所述第一数据矩阵中超过3σ1的元素的位置。Step S203, setting the first fault matrix R1 to record the positions of elements exceeding 3σ1 in thefirst data matrix.
在实际工程中，经常运用3σ准则的概念来确定随机变量的置信水平。即随机变量服从高斯分布：In practical engineering, the concept of 3σ criterion is often used to determine the confidence level of random variables. That is, the random variable obeys a Gaussian distribution:
数据落在(μ-σ,μ+σ,)中置信水平为68.26％；The confidence level of the data falling in (μ-σ,μ+σ,) is 68.26%;
数据落在(μ-2σ,μ+2σ,)中置信水平为95.44％；The data falls in (μ-2σ, μ+2σ,) with a confidence level of 95.44%;
数据落在(μ-3σ,μ+3σ,)中置信水平为99.74％；The data falls in (μ-3σ, μ+3σ,) with a confidence level of 99.74%;
由此可知，Y的取值几乎全部集中在(μ-3σ,μ+3σ,)区间范围内，如果数据落在±3σ之外，则认为是小概率事件，如果数组中出现某个单元大量数据落在±3σ之，则认为该单元受其他因素影响或单元存在故障。It can be seen that the values of Y are almost all concentrated in the interval range of (μ-3σ, μ+3σ,). If the data falls outside ±3σ, it is considered as a small probability event. If there is a large number of certain units in the array If the data falls within ±3σ, it is considered that the unit is affected by other factors or the unit is faulty.
计算(μ1-3σ1,μ1+3σ1)的取值范围，判断Uti.j值是否超过该取值范围，如果超过该取值范围，则设置第一故障矩阵R1记录超过该范围的元素的位置，并将该位置记为1，其余位置记为0。Calculate the value range of (μ1 -3σ1 , μ1 +3σ1 ), judge whether the value of Uti.j exceeds this value range, if it exceeds this value range, set the first fault matrix R1 record to exceed this range The position of the element of , and record this position as 1, and record the rest as 0.
步骤S204、剔除所述第一数据矩阵中第一故障矩阵R1记录的对应位置的元素获得第二数据矩阵，再重新计算第二数据矩阵的行向量的第二均值μ2和第二标准差σ2。Step S204, get rid of the element of the corresponding position recorded in the first failure matrix R1 in the first data matrix to obtain the second data matrix, and then recalculate the second mean value μ2 and the second standard deviation σ of the row vector of the second data matrix2 .
先将第一数据矩阵中超过3σ1的元素全部剔除，然后将剩余的数据重新放置第二数据矩阵中，得到第二数据矩阵为1行n列，最后通过数值计算工具计算第二数据矩阵的第二均值μ2和第二标准差σ2。First remove all the elements exceeding 3σ1 in the first data matrix, and then re-place the remaining data in the second data matrix to obtain the second data matrix with 1 row and n columns, and finally use the numerical calculation tool to calculate the value of the second data matrix Second mean value μ2 and second standard deviation σ2 .
步骤S205、设置第二故障矩阵R2记录所述第一数据矩阵中超过3σ2的元素的位置。Step S205, setting the second fault matrix R2 to record the positions of elements exceeding 3σ2 in the first data matrix.
同步骤203，可设置第二故障矩阵R2标记超过(μ2-3σ2,μ2+3σ2)的取值范围的元素的位置，将该位置记为1，其余位置记为0。Similar to step 203, the second fault matrix R2 can be set to mark the positions of elements exceeding the value range of (μ2 -3σ2 , μ2 +3σ2 ), and mark this position as 1, and mark the other positions as 0.
步骤S206、判断σ1-σ2的差值是否大于第一给定值，是，则μ1＝μ2，σ1＝σ2，R1＝R2，并返回到步骤S204；否，则执行步骤S207。Step S206, judging whether the difference between σ1 -σ2 is greater than the first given value, if yes, then μ1 = μ2 , σ1 =σ2 , R1 = R2, and return to step S204; if not, execute step S206 S207.
步骤S207、输出第二数据矩阵的行向量的第二均值μ2和第二标准差σ2，建立第二高斯分布。Step S207, output the second mean value μ2 and the second standard deviation σ2 of the row vectors of the second data matrix, and establish the second Gaussian distribution.
将上述的第二均值μ2和第二标准差σ2带入高斯分布函数即可绘制第二高斯分布曲线图，其服从正态分布(μ2，σ22)。Bringing the above-mentioned second mean value μ2 and second standard deviation σ2 into the Gaussian distribution function can draw the second Gaussian distribution curve, which obeys normal distribution (μ2 , σ22 ).
步骤S208、将第二故障矩阵R2中的每一列向量相加得到每个电池单体的故障数。Step S208, adding up each column vector in the second failure matrix R2 to obtain the number of failures of each battery cell.
其中，mti,j表示ti时刻第j个单体是否超过3σ2，超过3σ2，则将该位置记为1，其余位置记为0，即mti,j为0或1，然后将每一列的单体值进行叠加得到每个电池单体的故障数，即其中，1≤j≤n。 Among them, mti,j indicates whether the jth monomer exceeds 3σ2 at time ti, and if it exceeds 3σ2 , this position is recorded as 1, and other positions are recorded as 0, that is, mti,j is 0 or 1, and then each The cell values of one column are superimposed to obtain the number of failures of each battery cell, namely Among them, 1≤j≤n.
步骤S3、根据所述故障数计算每个电池单体的故障概率。Step S3, calculating the failure probability of each battery cell according to the number of failures.
其具体的计算公式为：m为所述数据矩阵的行数。因为研究数据矩阵是在大数据条件下进行，数据量非常大，所以根据大数定律，本申请还可以用故障频率表示故障概率，即Its specific calculation formula is: m is the row number of the data matrix. Because the research data matrix is carried out under the condition of big data, the amount of data is very large, so according to the law of large numbers, this application can also use the failure frequency to represent the failure probability, namely
步骤S4、判断每个电池单体的故障概率是否超过第二给定值，是，则确定故障发生，其具体步骤如下：Step S4, judging whether the failure probability of each battery cell exceeds the second given value, if yes, then determine that a failure has occurred, and the specific steps are as follows:
S401、根据所述故障概率判断是否超过第二给定值，是，则执行步骤S402，否，则执行步骤S403。S401. Determine whether the failure probability exceeds a second given value, if yes, execute step S402, and if no, execute step S403.
第二给定值可根据具体情况需要进行设置，可实现对故障值的筛选做到精确诊断和控制，为后期车辆数目、车辆种类、不同时间等多维度的大数据处理，提供了高效的算法。The second given value can be set according to the needs of the specific situation, which can realize the screening of fault values to achieve accurate diagnosis and control, and provide an efficient algorithm for multi-dimensional big data processing such as the number of vehicles, vehicle types, and different times in the later period .
S402、对电池单体进行报错。S402. Report an error to the battery cell.
判断故障概率是否超过第二给定值，如果超过第二给定值，则对电池单体进行报错。It is judged whether the failure probability exceeds the second given value, and if it exceeds the second given value, an error is reported to the battery cell.
S403、对超过第二给定值的电池单体统计规律，并得到相应的结果。S403. Statistical rules for battery cells exceeding the second given value, and obtain corresponding results.
判断故障概率是否超过第二给定值，如果没有超过第二给定值，对超过第二给定值的电池单体统计规律，并得到相应的结果。Judging whether the failure probability exceeds the second given value, if not exceeding the second given value, counting the battery cells exceeding the second given value, and obtaining corresponding results.
本发明实施例还提供了一种电力驱动交通工具的故障诊断系统，参见图3所示，包括获取模块1、3σ筛选模块2、求取故障概率模块3、故障判断模块4。The embodiment of the present invention also provides a fault diagnosis system for an electric drive vehicle, as shown in FIG.
获取模块1，用于获取被监控交通工具的电池的运行参数数据，对所述运行参数数据进行筛选整理得到被监控交通工具的电池单体的电压的第一数据矩阵。The acquisition module 1 is used to acquire the operating parameter data of the battery of the monitored vehicle, and filter and arrange the operating parameter data to obtain a first data matrix of the voltage of the battery cells of the monitored vehicle.
3σ筛选模块2，用于对所述第一数据矩阵采用3σ多层次筛选算法得到每个电池单体的故障数。The 3σ screening module 2 is configured to use a 3σ multi-level screening algorithm on the first data matrix to obtain the number of failures of each battery cell.
求取故障概率模块3，用于根据所述故障数计算每个电池单体的故障概率。Obtaining the failure probability module 3, used to calculate the failure probability of each battery cell according to the failure number.
故障判断模块4，用于判断每个电池单体的故障概率是否超过第二给定值，是，则确定故障发生。The failure judging module 4 is used to judge whether the failure probability of each battery cell exceeds a second given value, and if yes, determine that a failure has occurred.
3σ筛选模块2，具体用于：判断m行数据矩阵是否已完成循环，否，则计算第一数据矩阵的行向量的第一均值μ1和第一标准差σ1，建立第一高斯分布，记录所述第一数据矩阵中超过3σ1的元素的位置，剔除所述第一数据矩阵中第一故障矩阵R1记录的对应位置的元素获得第二数据矩阵，再重新计算第二数据矩阵的行向量的第二均值μ2和第二标准差σ2，记录所述第一数据矩阵中超过3σ2的元素的位置，直至σ1-σ2的差值是小于第一给定值为止，输出第二数据矩阵的行向量的第二均值μ2和第二标准差σ2，建立第二高斯分布，当m行数据矩阵已完成循环时，将第二故障矩阵R2中的每一列向量相加得到每个电池单体的故障数。The 3σ screening module 2 is specifically used for: judging whether the m-row data matrix has completed the cycle; if not, then calculate the first mean value μ1 and the first standard deviation σ1 of the row vector of the first data matrix, and establish the first Gaussian distribution, Record the positions of elements exceeding 3σ1 in thefirst data matrix, remove the elements in the corresponding positions recorded by the first failure matrix R1 in the first data matrix to obtain the second data matrix, and then recalculate the rows of the second data matrix The second mean value μ2 and the second standard deviation σ2 of the vector, record the positions of elements exceeding 3σ2 in the first data matrix, until the difference between σ1 -σ2 is less than the first given value, output the first The second mean value μ2 and the second standard deviation σ2 of the row vectors of the second data matrix establish a second Gaussian distribution, and when the m-row data matrix has completed the cycle, add each column vector in the second fault matrix R2 to obtain The number of failures per battery cell.
求取故障概率模块3，用于采用如下式子计算每个电池单体的故障概率：m为所述数据矩阵的行数。Obtaining the failure probability module 3 is used to calculate the failure probability of each battery cell using the following formula: m is the number of rows of the data matrix.
故障判断模块4包括第二逻辑判断单元、故障报错单元、统计规律单元。The fault judgment module 4 includes a second logic judgment unit, a fault error reporting unit, and a statistical law unit.
第二逻辑判断单元，用于根据所述故障概率判断是否超过第二给定值。The second logic judging unit is configured to judge whether the failure probability exceeds a second given value according to the failure probability.
故障报错单元，用于对电池单体进行报错。The fault error reporting unit is used to report errors to the battery cells.
统计规律单元，用于对超过第二给定值的电池单体统计规律，并得到相应的结果。The statistical law unit is used for statistical law of the battery cells exceeding the second given value, and obtains corresponding results.
为了进一步更好的论述3σ多层次筛选算法的内容，下面举例进行说明：In order to further and better discuss the content of the 3σ multi-level screening algorithm, the following example is used to illustrate:
以北京电动汽车在2016年4月11日行驶的某一时刻采集的电池的单体电压为例，采用多层次3σ筛选算法对采集的91个电池的单体电压进行诊断，所以构成的第一数据矩阵为1行91列，即第一数据矩阵的电压为:Taking the battery cell voltage collected by Beijing electric vehicles at a certain moment on April 11, 2016 as an example, the multi-level 3σ screening algorithm was used to diagnose the collected 91 battery cell voltages, so the first The data matrix is 1 row and 91 columns, that is, the voltage of the first data matrix is:
U0＝U0 =
[[
4.024.024.024.024.024.0944.024.0244.024.024.024.024.024.024.024.0244.024.0244.024.024.024.024.024.02444.024.024.024.024.024.024.0244.024.024.14.024.024.024.024.024.024.0244.024.024.024.024.024.024.024.024.024.024.024.024.024.024.024.024.0244.024.024.024.024.024.024.024.024.0244.024.024.024.024.024.024.024.024.024.024.024.024.024.024.024.024.024.024.024.0944.024.0244.024.024.024.024.024.024.024.0244.024.0244.024.024.024.024.024.02444.024.024.024.024.024.024.0244.024.024.14.024.024.024.024.024.024.0244.024.024.024.024.024.024.024.024. 024.024.024.024.024.024.024.024.0244.024.024.024.024.024.024.024.024.0244.024.024.024.024.024.024.024.024.024.024.024.024.024.
]]
将第一数据矩阵的数据输入值到已编好的程序中计算第一均值μ1和第一标准差σ1，即第一数据矩阵的第一均值为μ1＝4.0195，第一标准差为σ1＝0.0130。根据第一数据矩阵可绘制出图4电池的单体电压曲线图；根据第一均值为μ1，第一标准差为σ1可绘制出图5第一高斯分布曲线图。Put the data input value of the first data matrix into the compiled program to calculate the first mean value μ1 and the first standard deviation σ1 , that is, the first mean value of the first data matrix is μ1 =4.0195, and the first standard deviation is σ1 =0.0130. According to the first data matrix, the cell voltage curve of the battery in Fig. 4 can be drawn; according to the first mean value of μ1 and the first standard deviation of σ1 , the first Gaussian distribution curve of Fig. 5 can be drawn.
由图4可看出，第6位置和第41位置电压值明显超出电池组电压平均水平，因为计算这组端电压的均值时不希望被这两个位置影响，所以在计算均值时，利用3σ多层次筛选算法将这两个值剔除出去。It can be seen from Figure 4 that the voltage values at the 6th position and the 41st position obviously exceed the average level of the battery pack voltage, because the average value of the terminal voltage of this group is not expected to be affected by these two positions, so when calculating the average value, use 3σ The multi-level filtering algorithm removes these two values.
根据3σ多层次数据筛选算法，在(μ1-3σ1,μ1+3σ1)区间之外则为第一故障值，因此3σ1的上限为：According to the 3σ multi-level data screening algorithm, it is the first fault value outside the interval (μ1 -3σ1 , μ1 +3σ1 ), so the upper limit of 3σ1 is:
μ1+3σ1＝4.0195+3×0.0130＝4.0585μ1 +3σ1 =4.0195+3×0.0130=4.0585
3σ1的下限为：The lower bound of 3σ1 is:
μ1-3σ1＝4.0195-3×0.0130＝3.9805μ1 -3σ1 = 4.0195-3×0.0130 = 3.9805
3σ1为：(3.9805,4.0585)3σ1 is: (3.9805,4.0585)
显然，第6位置和第41位置的值明显超出3σ1，所以将第一故障矩阵R1矩阵中第6位置和第41位置为1，其余位置为0，剔除所述第一数据矩阵中第一故障矩阵R1记录的对应位置的元素获得第二数据矩阵，此时第二数据矩阵U2为1行91列，重新将第二数据矩阵导入数值计算工具中计算第二均值μ2和第二标准差σ2，即第二均值μ2＝4.0178、第二标准差σ2＝0.0064，3σ2为(3.9986,4.037)，显然剔除后的数据全部落在此区间内，实现了一次筛选能同时去除掉所有超过3σ2的数据，处理效率高。Obviously, the values of the 6th position and the 41st position obviously exceed 3σ1 , so the 6th position and the 41st position in the first fault matrix R1 matrix are set to 1, and the remaining positions are 0, and the first The element of the corresponding position recorded in the fault matrix R1 obtains the second data matrix, at this time, the second data matrix U2 is 1 row and 91 columns, and the second data matrix is imported into the numerical calculation tool to calculate the second mean value μ2 and the second standard The difference σ2 is the second mean value μ2 = 4.0178, the second standard deviation σ2 = 0.0064, and the 3σ2 is (3.9986, 4.037). Obviously, all the data after elimination fall within this range, realizing that one screening can simultaneously remove All data exceeding 3σ2 can be processed with high efficiency.
根据第二均值μ2和第二标准差σ2绘制出图6第二高斯分布曲线图，为了便于研究第一高斯分布曲线图和第二高斯分布曲线图的区别，将两个高斯分布图输出到同一张图纸上，即图7为两次高斯分布比较曲线图。According to the second mean value μ2 and the second standard deviation σ 2 draw the second Gaussian distribution curve figure in Fig. 6, in order to facilitate the research on the difference between the first Gaussian distribution curve figure and the second Gaussian distribution curve figure, two Gaussian distribution curve figures are exported to On the same drawing, that is, Figure 7 is a comparison curve of two Gaussian distributions.
由图7可知，去掉个别超出3σ1范围的数据后，第二标准差下降到第一标准差的49.2％。第一均值μ1与第二均值μ2之间相差一个差值，这个差值就代表了一次筛选后对原来中心值进行的调整。从第一数据矩阵中可知，大部分数据都落在4.00和4.02之间，因此第一均值大约在4.01左右，但因为第6和第41数据异常偏大，导致第一均值为4.019，因此去掉个别超出3σ1取值范围的数据后得到的中心值将更加准确地确定该无故障系统整体的中心位置，便于在高斯分布中定位故障或异常的位置及置信水平。It can be seen from Figure 7 that after removing individual data beyond the range of 3σ1 , the second standard deviation drops to 49.2% of the first standard deviation. There is a difference between the first mean value μ1 and the second mean value μ2 , and this difference represents an adjustment to the original central value after one screening. It can be seen from the first data matrix that most of the data fall between 4.00 and 4.02, so the first mean value is about 4.01, but because the 6th and 41st data are abnormally large, the first mean value is 4.019, so remove The central value obtained from individual data beyond the value range of 3σ1 will more accurately determine the central position of the fault-free system as a whole, which is convenient for locating the position and confidence level of the fault or abnormality in the Gaussian distribution.
采用多层次3σ筛选算法具有以下优点：一、一次筛选能同时去除掉所有超过3σ置信区间的数据，处理效率高；二、能按照置信区间来调整第二给定值，为后期车辆数目、车辆种类、不同时间等多维度的大数据处理，提供了高效的算法；三、用第二故障矩阵R2将超3σ的数据位置保存，便于后期对电池组组内电池单体之间故障规律进行预处理。The multi-level 3σ screening algorithm has the following advantages: 1. One screening can remove all the data exceeding the 3σ confidence interval at the same time, and the processing efficiency is high; 2. The second given value can be adjusted according to the confidence interval, which is the number of vehicles, vehicles Multi-dimensional big data processing such as types and different times provides an efficient algorithm; 3. Use the second fault matrix R2 to save the data position exceeding 3σ, which is convenient for later prediction of the fault law between the battery cells in the battery pack deal with.
本说明书中各个实施例采用递进的方式描述，每个实施例重点说明的都是与其他实施例的不同之处，各个实施例之间相同相似部分互相参见即可。对于实施例公开的系统而言，由于其与实施例公开的方法相对应，所以描述的比较简单，相关之处参见方法部分说明即可。Each embodiment in this specification is described in a progressive manner, each embodiment focuses on the difference from other embodiments, and the same and similar parts of each embodiment can be referred to each other. As for the system disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and for the related information, please refer to the description of the method part.
本文中应用了具体个例对本发明的原理及实施方式进行了阐述，以上实施例的说明只是用于帮助理解本发明的方法及其核心思想；同时，对于本领域的一般技术人员，依据本发明的思想，在具体实施方式及应用范围上均会有改变之处。综上所述，本说明书内容不应理解为对本发明的限制。In this paper, specific examples have been used to illustrate the principle and implementation of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea; meanwhile, for those of ordinary skill in the art, according to the present invention Thoughts, there will be changes in specific implementation methods and application ranges. In summary, the contents of this specification should not be construed as limiting the present invention.
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610834850.0ACN106383951B (en) | 2016-09-20 | 2016-09-20 | Fault diagnosis method and system for an electric drive vehicle |
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610834850.0ACN106383951B (en) | 2016-09-20 | 2016-09-20 | Fault diagnosis method and system for an electric drive vehicle |
| Publication Number | Publication Date |
|---|---|
| CN106383951A CN106383951A (en) | 2017-02-08 |
| CN106383951Btrue CN106383951B (en) | 2019-08-06 |
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610834850.0AActiveCN106383951B (en) | 2016-09-20 | 2016-09-20 | Fault diagnosis method and system for an electric drive vehicle |
| Country | Link |
|---|---|
| CN (1) | CN106383951B (en) |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109101390B (en)* | 2018-06-29 | 2021-08-24 | 平安科技（深圳）有限公司 | Timed task abnormity monitoring method based on Gaussian distribution, electronic device and medium |
| CN112115618A (en)* | 2020-09-22 | 2020-12-22 | 南方电网海南数字电网研究院有限公司 | Power equipment fault diagnosis method and system based on matrix chart and confidence |
| CN112986839B (en)* | 2021-02-25 | 2021-12-03 | 北京理工大学 | Confidence interval-based fault diagnosis method and system for lithium ion power battery pack |
| CN113868236B (en)* | 2021-09-29 | 2023-06-30 | 东风汽车有限公司东风日产乘用车公司 | Electric vehicle fault monitoring method, equipment and computer readable storage medium |
| CN118672237A (en)* | 2024-06-25 | 2024-09-20 | 比亚迪股份有限公司 | Fault analysis method, system, storage medium, electronic device, and program product |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6330362B1 (en)* | 1996-11-12 | 2001-12-11 | Texas Instruments Incorporated | Compression for multi-level screened images |
| CN103057433A (en)* | 2012-12-25 | 2013-04-24 | 北京理工大学 | Method for detecting unit cells of grouped cells in electric automobile under working state |
| CN203039043U (en)* | 2013-01-08 | 2013-07-03 | 北京优科利尔能源设备有限公司 | Detection and maintenance system for power battery of electric vehicle |
| CN103986625A (en)* | 2014-05-29 | 2014-08-13 | 中国科学院软件研究所 | A Cloud Application Fault Diagnosis System Based on Statistical Monitoring |
| CN104021238A (en)* | 2014-03-25 | 2014-09-03 | 重庆邮电大学 | Lead-acid power battery system fault diagnosis method |
| CN104462802A (en)* | 2014-11-26 | 2015-03-25 | 浪潮电子信息产业股份有限公司 | Method for analyzing outlier data in large-scale data |
| CN105372600A (en)* | 2015-11-24 | 2016-03-02 | 珠海朗尔电气有限公司 | Storage battery fault diagnosis system and method based on big data |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7715961B1 (en)* | 2004-04-28 | 2010-05-11 | Agnik, Llc | Onboard driver, vehicle and fleet data mining |
| EP2951722B1 (en)* | 2013-01-31 | 2018-05-16 | Universite De Montpellier | Process for identifying rare events |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6330362B1 (en)* | 1996-11-12 | 2001-12-11 | Texas Instruments Incorporated | Compression for multi-level screened images |
| CN103057433A (en)* | 2012-12-25 | 2013-04-24 | 北京理工大学 | Method for detecting unit cells of grouped cells in electric automobile under working state |
| CN203039043U (en)* | 2013-01-08 | 2013-07-03 | 北京优科利尔能源设备有限公司 | Detection and maintenance system for power battery of electric vehicle |
| CN104021238A (en)* | 2014-03-25 | 2014-09-03 | 重庆邮电大学 | Lead-acid power battery system fault diagnosis method |
| CN103986625A (en)* | 2014-05-29 | 2014-08-13 | 中国科学院软件研究所 | A Cloud Application Fault Diagnosis System Based on Statistical Monitoring |
| CN104462802A (en)* | 2014-11-26 | 2015-03-25 | 浪潮电子信息产业股份有限公司 | Method for analyzing outlier data in large-scale data |
| CN105372600A (en)* | 2015-11-24 | 2016-03-02 | 珠海朗尔电气有限公司 | Storage battery fault diagnosis system and method based on big data |
| Title |
|---|
| Model-based fault detection of a battery system in a hybrid electric vehicle;S.A. Gadsden等;《2011 IEEE Vehicle Power and Propulsion Conference》;20111013;第1-6页 |
| 车用动力锂离子电池系统故障诊断研究与实现;檀斐;《中国优秀硕士学位论文全文数据库 工程科技Ⅱ辑》;20150715(第07期);第C035-412页 |
| Publication number | Publication date |
|---|---|
| CN106383951A (en) | 2017-02-08 |
| Publication | Publication Date | Title |
|---|---|---|
| CN106383951B (en) | Fault diagnosis method and system for an electric drive vehicle | |
| CN106353690B (en) | Utilize the method for Petri network diagnosis lithium battery failure | |
| CN113406524B (en) | A method and system for inconsistent fault diagnosis of power battery system | |
| CN115166555B (en) | Method and system for identifying abnormal internal resistance of battery module of battery big data platform | |
| CN111579121A (en) | Method for diagnosing temperature fault in new energy automobile battery pack on line based on big data | |
| CN117607704A (en) | Lithium ion battery pack micro-short circuit fault diagnosis method considering inconsistency | |
| CN118858961B (en) | Lithium ion battery micro-short circuit diagnosis method and system based on clustering and charging voltage difference | |
| CN115128468A (en) | Chemical energy storage battery PHM undervoltage fault prediction method | |
| CN118566732A (en) | Fault diagnosis method and system based on outlier measurement of electromechanical signals of lithium-ion batteries | |
| CN111562503B (en) | A method for analyzing and processing lithium-ion battery charging and discharging equipment that does not fail | |
| CN117572241A (en) | Lithium battery fault analysis method based on extended multidimensional time sequence | |
| CN118708882A (en) | Multi-source data collection and aggregation method, system, device and storage medium | |
| CN119749240A (en) | Power battery abnormality diagnosis method and device for new energy automobile and storage medium | |
| CN119471393B (en) | A vehicle power battery fault diagnosis method based on hierarchical identification | |
| CN103196691B (en) | Method for establishing relevance fault testing matrix based on main diagonal element domination fuzzy relation matrix | |
| WO2022162060A1 (en) | Big data for fault identification in battery systems | |
| CN118604640A (en) | Battery evaluation method, device, electronic device, storage medium and program product | |
| DE102021006561A1 (en) | Big data for error detection in battery systems | |
| CN115061043B (en) | A method, device and electronic device for screening defective battery cells | |
| CN117031294A (en) | Battery multi-fault detection method, device and storage medium | |
| CN113405743B (en) | New energy electric vehicle production and manufacturing test data analysis processing method and system based on cloud computing and storage medium | |
| CN115981945A (en) | Nuclear power instrument control board card based hardware health parameter verification method and device | |
| CN115902631A (en) | Power battery fault detection method and device | |
| KR20230026615A (en) | Architecture for measurement system designed for massive impedance extraction of battery systems | |
| CN115608653A (en) | Method and device for screening abnormal battery units in battery system |
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |