1.数据集介绍
将故障区分为具体的不同类型：单相短路故障、两相接地短路故障、两相相间故障、三相相间短路故障。这里随意举出每种类别的两个样本进行展示。
| G | C | B | A | Ia | Ib | Ic | Va | Vb | Vc |
| 1 | 0 | 0 | 1 | -151.2918124 | -9.677451563 | 85.80016226 | 0.400749853 | -0.132934945 | -0.267814907 |
| 1 | 0 | 0 | 1 | -336.1861826 | -76.28326195 | 18.32889658 | 0.312731934 | -0.123633156 | -0.189098779 |
| 1 | 0 | 1 | 1 | -343.4870147 | 104.5627513 | 3.794285309 | 0.272042501 | 0.011317575 | -0.283360076 |
| 1 | 0 | 1 | 1 | -339.1254001 | 105.4293167 | -0.267241225 | 0.27782054 | 0.021756839 | -0.299577378 |
| 0 | 1 | 1 | 0 | 19.38615173 | -785.553797 | 768.7279081 | -0.210406869 | -0.0020112 | 0.212418069 |
| 0 | 1 | 1 | 0 | 18.47841651 | -783.8619173 | 767.9410527 | -0.217651204 | -0.00260451 | 0.220255714 |
| 0 | 1 | 1 | 1 | 506.5917463 | 374.8825788 | -879.344997 | 0.042029705 | -0.025636401 | -0.016393305 |
| 0 | 1 | 1 | 1 | 495.1384715 | 387.4159615 | -880.4253096 | 0.042107683 | -0.025103056 | -0.017004627 |
2.模型：BP网络，每类故障有1000个样本
3.效果（平均识别准确率为95.304%）
混淆矩阵
F1-score
误差曲线
4.代码
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
import warnings
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import LabelEncoder
from sklearn.neural_network import MLPClassifier
from sklearn.metrics import accuracy_score, classification_report, confusion_matrix
sns.set_style('darkgrid')
plt.rcParams['figure.figsize'] = (13, 9)
plt.rcParams['font.size'] = 20
warnings.filterwarnings('ignore')
plt.rcParams['font.family'] = 'SimHei' # 设置字体为黑体
multi_data = pd.read_csv('./Data1.csv').iloc[0:4362, :]
lines = [
('Ia', 'Va'),
('Ib', 'Vb'),
('Ic', 'Vc')
]
#可以关注：https://mbd.pub/o/bread/ZJuXmZhv
print(f'Number of Samples: {multi_data.shape[0]}')
multi_data['faultType'] = multi_data['G'].astype(str) + multi_data['C'].astype(str) + multi_data['B'].astype(str) + \
multi_data['A'].astype(str)
multi_data.head()
