Analog Circuits

Each file, named File nameCircuit.txt is used for training and testing a system (SVM in our case) which performs the classification of the faults. Therefore It contains a set of data formatted for being the input of a classification software, and containing the following lines: n1 n2 varying components: c1 c2 ...... cNC mag_1,1 pha_1,1 mag_1,2 pha_1,2 ........ mag_1,nf pha1,nf FC_1 mag_2,1 pha_2,1 mag_2,2 pha_2,2 ........ mag_2,nf pha2,nf FC_2 ............. mag_Ne,1 phaNe,1 mag_Ne,2 pha_Ne,2 ........ mag_Ne,nf pha_Ne,nf FC_Ne ------------------- The exact meaning of any element is: n1 = number of input lines (dimension of the input vector) n2 = number of output lines (for this application = 1, because it is a classification problem and we have just one output integer which assumes the value of various classes) mag_i,j = it is the magnitude of i-th sample of the frequency response (the absolute value of the complex frequency response term at that given frequency) mag_i,j = it is the phase in rad of i-th sample of the frequency response (the phase of the complex frequency response term at that given frequency, that arctan(ImF/ReF) in rad) nf = it is the numer of sampling frequencies Ne = it is the number of examples FC_i = it is the correct Fault Class for that sample; the class 0 is "no faulty", otherwise it is an integer which represents the component which is out of tolerance (or, in case of ambiguity groups with a dim > 1, it represents an ambiguity group). as consequence to previous format we can also say that: n1 = 2 x nf (magnitude and phase each sample) mag_i,j are abs values with no normalization in the file and so they are always positive pha_i,j are the original phases in rad, so they are values ranging in the interval [-pi pi], not normalized in the file

每个名为`File nameCircuit.txt`的文件均用于训练与测试用于故障分类的系统（本研究中采用支持向量机（Support Vector Machine，SVM））。该文件包含一组适配分类软件输入格式的数据集，其结构如下：
`n1 n2 varying components: c1 c2 ...... cNC mag_1,1 pha_1,1 mag_1,2 pha_1,2 ........ mag_1,nf pha1,nf FC_1 mag_2,1 pha_2,1 mag_2,2 pha_2,2 ........ mag_2,nf pha2,nf FC_2 ............. mag_Ne,1 phaNe,1 mag_Ne,2 pha_Ne,2 ........ mag_Ne,nf pha_Ne,nf FC_Ne -------------------`
各元素的具体含义如下：
- n1：输入线路数（即输入向量的维度）
- n2：输出线路数（本应用场景中取值为1，因本任务为分类问题，仅需单个输出整数以表征不同故障类别）
- mag_i,j：第i个频响样本的幅值，即给定频率下复频响项的绝对值
- pha_i,j：第i个频响样本的相位（单位为弧度），即给定频率下复频响项的相位，计算公式为arctan(ImF/ReF)，单位为弧度
- nf：采样频率的总数量
- Ne：样本的总数量
- FC_i：对应样本的正确故障类别；其中类别0代表"无故障"，其余整数代表超出公差范围的元器件；若存在维度大于1的歧义组，则该值代表对应歧义组的编号。
基于上述格式，可进一步推导得到以下规则：
1. n1 = 2 × nf，即每个采样频率对应幅值与相位两类数据
2. 文件中存储的mag_i,j为未归一化的幅值，因此始终为正值
3. 文件中存储的pha_i,j为原始相位值，取值范围为[-π, π]，未经过归一化处理。

搜集汇总

背景与挑战

背景概述

Analog Circuits数据集是一个用于模拟电路故障分类的数据集，包含多个文本文件，每个文件包含频率响应的幅度和相位信息以及故障类别标签。该数据集适用于电气工程和系统故障诊断领域的研究。