个人工具合集
Project description
10分钟快速开始
前言
一些常用的工具函数，主要为客车整车试验数据处理相关内容。本文仅介绍了部分常用函数，data模块中还有许多函数没有涉及，详细用法请参考源码中的函数说明。文中出现的所有演示素材都可以在库安装路径下的examples文件夹中找到。
数据处理基本操作
读取表格文件
目前只封装了xlsx、xls和csv格式的文件读取函数，其他pandas支持的文件格式可以直接使用pandas内置的函数读取：
import xyw_utils as xy
processor = xy.data.FileProcessorFacade()
df = processor.read('CCBC.csv')
>>> df
Time [s] uDCDCIn Unnamed: 2 Time [s].1 ... Time [s].11 rBattSOC Unnamed: 35 Unnamed: 36
0 0.01149 617.0 NaN 0.01207 ... 0.09046 97.6 NaN NaN
1 4.01241 617.0 NaN 4.01299 ... 0.19049 97.6 NaN NaN
2 5.01263 617.0 NaN 5.01320 ... 0.29054 97.6 NaN NaN
3 6.01286 617.0 NaN 6.01344 ... 3.49059 97.6 NaN NaN
4 7.01309 617.0 NaN 7.01367 ... 3.59036 97.6 NaN NaN
... ... ... ... ... ... ... ... ... ...
13265 NaN NaN NaN NaN ... 1329.64510 96.4 NaN NaN
13266 NaN NaN NaN NaN ... 1329.74509 96.4 NaN NaN
13267 NaN NaN NaN NaN ... 1329.84511 96.4 NaN NaN
13268 NaN NaN NaN NaN ... 1329.94501 96.4 NaN NaN
13269 NaN NaN NaN NaN ... NaN NaN NaN NaN
[13270 rows x 37 columns]
对于只有单一时间序列的数据，使用上述命令就可以直接生成以表格中第一行为列名的pandas.DataFrame数据，不过对于每一信号都有单独时间序列的数据则需要继续进行如下操作：
dfs = xy.data.init_split(df)
>>> dfs
[ Time [s] uDCDCIn
0 0.01149 617.0
1 4.01241 617.0
2 5.01263 617.0
3 6.01286 617.0
4 7.01309 617.0
... ... ...
1322 1325.31101 617.0
1323 1326.31125 617.0
1324 1327.31149 617.0
1325 1328.31172 617.0
1326 1329.31193 617.0
[1327 rows x 2 columns], Time [s] iDCDCIn
0 0.01207 2.0
1 4.01299 3.0
2 5.01320 2.0
3 6.01344 3.0
4 7.01367 2.0
... ... ...
1322 1325.31158 3.0
1323 1326.31183 2.0
1324 1327.31207 2.0
1325 1328.31229 2.0
1326 1329.31251 2.0
[1327 rows x 2 columns], Time [s] iMCU
0 0.02674 2.0
1 0.12668 2.0
2 0.22669 2.0
3 3.52688 2.0
4 3.62686 2.0
... ... ...
13263 1329.59782 3.0
13264 1329.69795 3.0
13265 1329.79783 3.0
13266 1329.89784 3.0
13267 1329.99790 3.0
[13268 rows x 2 columns], Time [s] uMCU
0 0.02674 617.0
1 0.12668 617.0
2 0.22669 617.0
3 3.52688 617.0
4 3.62686 617.0
... ... ...
13263 1329.59782 617.0
13264 1329.69795 617.0
13265 1329.79783 617.0
13266 1329.89784 617.0
13267 1329.99790 617.0
[13268 rows x 2 columns], Time [s] vVehSpdAbs
0 0.02730 0.0
1 0.12725 0.0
2 0.22725 0.0
3 3.52744 0.0
4 3.62742 0.0
... ... ...
13263 1329.59837 0.0
13264 1329.69851 0.0
13265 1329.79839 0.0
13266 1329.89839 0.0
13267 1329.99846 0.0
[13268 rows x 2 columns], Time [s] iSPCOut
0 0.05502 1
1 0.15502 0
2 0.25507 1
3 3.45522 1
4 3.55571 1
... ... ...
13265 1329.56053 1
13266 1329.66052 1
13267 1329.76050 1
13268 1329.86048 1
13269 1329.96046 1
[13270 rows x 2 columns], Time [s] uSPCOut
0 0.05502 222
1 0.15502 222
2 0.25507 230
3 3.45522 222
4 3.55571 222
... ... ...
13265 1329.56053 223
13266 1329.66052 223
13267 1329.76050 223
13268 1329.86048 222
13269 1329.96046 222
[13270 rows x 2 columns], Time [s] iAPCOut
0 0.07561 2
1 0.17561 2
2 0.27562 2
3 3.47580 0
4 3.57581 0
... ... ...
13265 1329.60061 0
13266 1329.70075 0
13267 1329.80063 0
13268 1329.90063 0
13269 1330.00179 0
[13270 rows x 2 columns], Time [s] uAPCOut
0 0.07561 327
1 0.17561 327
2 0.27562 326
3 3.47580 0
4 3.57581 0
... ... ...
13265 1329.60061 0
13266 1329.70075 0
13267 1329.80063 0
13268 1329.90063 0
13269 1330.00179 0
[13270 rows x 2 columns], Time [s] iBattCurr
0 0.09046 3.6
1 0.19049 3.8
2 0.29054 3.4
3 3.49059 1.2
4 3.59036 1.2
... ... ...
13264 1329.54537 1.2
13265 1329.64510 1.2
13266 1329.74509 1.2
13267 1329.84511 1.2
13268 1329.94501 1.2
[13269 rows x 2 columns], Time [s] uBattVltg
0 0.09046 617.7
1 0.19049 617.7
2 0.29054 617.7
3 3.49059 617.9
4 3.59036 617.9
... ... ...
13264 1329.54537 617.9
13265 1329.64510 617.9
13266 1329.74509 617.9
13267 1329.84511 617.9
13268 1329.94501 617.9
[13269 rows x 2 columns], Time [s] rBattSOC
0 0.09046 97.6
1 0.19049 97.6
2 0.29054 97.6
3 3.49059 97.6
4 3.59036 97.6
... ... ...
13264 1329.54537 96.4
13265 1329.64510 96.4
13266 1329.74509 96.4
13267 1329.84511 96.4
13268 1329.94501 96.4
[13269 rows x 2 columns]]
df = xy.data.merge_all(dfs)
>>> df
Time [s] uDCDCIn iDCDCIn iMCU uMCU ... iAPCOut uAPCOut iBattCurr uBattVltg rBattSOC
0 0.01149 617.0 NaN NaN NaN ... NaN NaN NaN NaN NaN
1 0.01207 NaN 2.0 NaN NaN ... NaN NaN NaN NaN NaN
2 0.02674 NaN NaN 2.0 617.0 ... NaN NaN NaN NaN NaN
3 0.02730 NaN NaN NaN NaN ... NaN NaN NaN NaN NaN
4 0.05502 NaN NaN NaN NaN ... NaN NaN NaN NaN NaN
... ... ... ... ... ... ... ... ... ... ... ...
68994 1329.94501 NaN NaN NaN NaN ... NaN NaN 1.2 617.9 96.4
68995 1329.96046 NaN NaN NaN NaN ... NaN NaN NaN NaN NaN
68996 1329.99790 NaN NaN 3.0 617.0 ... NaN NaN NaN NaN NaN
68997 1329.99846 NaN NaN NaN NaN ... NaN NaN NaN NaN NaN
68998 1330.00179 NaN NaN NaN NaN ... 0.0 0.0 NaN NaN NaN
[68999 rows x 13 columns]
处理到这一步时，已经将各时间序列的数据统一到了同一个总的时间轴上，不过每一个信号列中都会有许多NaN的无效数据，可以使用下面的方法进行插值：
df = xy.data.interpolate_nan(df, drop=False)
>>> df
Time [s] uDCDCIn iDCDCIn iMCU uMCU ... iAPCOut uAPCOut iBattCurr uBattVltg rBattSOC
0 0.01149 617.0 NaN NaN NaN ... NaN NaN NaN NaN NaN
1 0.01207 617.0 2.000000 NaN NaN ... NaN NaN NaN NaN NaN
2 0.02674 617.0 2.003667 2.0 617.0 ... NaN NaN NaN NaN NaN
3 0.02730 617.0 2.003807 2.0 617.0 ... NaN NaN NaN NaN NaN
4 0.05502 617.0 2.010735 2.0 617.0 ... NaN NaN NaN NaN NaN
... ... ... ... ... ... ... ... ... ... ... ...
68994 1329.94501 NaN NaN 3.0 617.0 ... 0.0 0.0 1.2 617.9 96.4
68995 1329.96046 NaN NaN 3.0 617.0 ... 0.0 0.0 NaN NaN NaN
68996 1329.99790 NaN NaN 3.0 617.0 ... 0.0 0.0 NaN NaN NaN
68997 1329.99846 NaN NaN NaN NaN ... 0.0 0.0 NaN NaN NaN
68998 1330.00179 NaN NaN NaN NaN ... 0.0 0.0 NaN NaN NaN
[68999 rows x 13 columns]
注意：当把drop参数设为False时，首尾无法插值的部分数据依旧会保持为NaN。当确认不需要这部分数据时，可以直接将drop设为True，舍去该部分数据；也可以使用xy.data.remove_nan()方法手动删去所有带有NaN数据的行：
df = xy.data.remove_nan(df)
>>> df
Time [s] uDCDCIn iDCDCIn iMCU uMCU ... iAPCOut uAPCOut iBattCurr uBattVltg rBattSOC
0 0.09046 617.0 2.019593 2.0 617.0 ... 2.0 327.0 3.600000 617.7 97.6
1 0.12668 617.0 2.028646 2.0 617.0 ... 2.0 327.0 3.672418 617.7 97.6
2 0.12725 617.0 2.028788 2.0 617.0 ... 2.0 327.0 3.673558 617.7 97.6
3 0.15502 617.0 2.035729 2.0 617.0 ... 2.0 327.0 3.729081 617.7 97.6
4 0.17561 617.0 2.040876 2.0 617.0 ... 2.0 327.0 3.770249 617.7 97.6
... ... ... ... ... ... ... ... ... ... ... ...
68952 1329.26058 617.0 2.000000 3.0 617.0 ... 0.0 0.0 1.200000 617.9 96.4
68953 1329.29780 617.0 2.000000 3.0 617.0 ... 0.0 0.0 1.200000 617.9 96.4
68954 1329.29836 617.0 2.000000 3.0 617.0 ... 0.0 0.0 1.200000 617.9 96.4
68955 1329.30060 617.0 2.000000 3.0 617.0 ... 0.0 0.0 1.200000 617.9 96.4
68956 1329.31193 617.0 2.000000 3.0 617.0 ... 0.0 0.0 1.200000 617.9 96.4
[68957 rows x 13 columns]
保存数据
保存的格式与读取时的限制相同：
processor.write(df, 're.xlsx')
积分
封装了两种常用的离散数据积分方法，辛普森法和复杂梯形法，这里以积分电池能耗为例：
df['pb'] = df['iBattCurr'] * df['uBattVltg']
df['wb'] = xy.data.get_integral(df[['Time [s]', 'pb']], method='trapz', last_only=False) / 3600000
>>> df
Time [s] uDCDCIn iDCDCIn iMCU uMCU ... iBattCurr uBattVltg rBattSOC pb wb
0 0.09046 617.0 2.019593 2.0 617.0 ... 3.600000 617.7 97.6 2223.720000 0.000000
1 0.12668 617.0 2.028646 2.0 617.0 ... 3.672418 617.7 97.6 2268.452768 0.000023
2 0.12725 617.0 2.028788 2.0 617.0 ... 3.673558 617.7 97.6 2269.156735 0.000023
3 0.15502 617.0 2.035729 2.0 617.0 ... 3.729081 617.7 97.6 2303.453504 0.000041
4 0.17561 617.0 2.040876 2.0 617.0 ... 3.770249 617.7 97.6 2328.882761 0.000054
... ... ... ... ... ... ... ... ... ... ... ...
68952 1329.26058 617.0 2.000000 3.0 617.0 ... 1.200000 617.9 96.4 741.480000 2.526091
68953 1329.29780 617.0 2.000000 3.0 617.0 ... 1.200000 617.9 96.4 741.480000 2.526098
68954 1329.29836 617.0 2.000000 3.0 617.0 ... 1.200000 617.9 96.4 741.480000 2.526099
68955 1329.30060 617.0 2.000000 3.0 617.0 ... 1.200000 617.9 96.4 741.480000 2.526099
68956 1329.31193 617.0 2.000000 3.0 617.0 ... 1.200000 617.9 96.4 741.480000 2.526101
[68957 rows x 15 columns]
注意：method参数只支持‘simps’以及‘trapz’两种，last_only设为True时，只返回最终积分结果，可以大大减少运算时间，此例中即2.526101。对于求电量积分这一操作，也可以使用xy.vehicle_test.energy_consumption_test()这一专用函数：
df['wb'] = xy.vehicle_test.energy_consumption_test(df, 1, 'Time [s]', 'iBattCurr', 'uBattVltg', last_only=False)
>>> df
Time [s] uDCDCIn iDCDCIn iMCU uMCU ... iBattCurr uBattVltg rBattSOC pb wb
0 0.09046 617.0 2.019593 2.0 617.0 ... 3.600000 617.7 97.6 2223.720000 0.000000
1 0.12668 617.0 2.028646 2.0 617.0 ... 3.672418 617.7 97.6 2268.452768 0.000023
2 0.12725 617.0 2.028788 2.0 617.0 ... 3.673558 617.7 97.6 2269.156735 0.000023
3 0.15502 617.0 2.035729 2.0 617.0 ... 3.729081 617.7 97.6 2303.453504 0.000041
4 0.17561 617.0 2.040876 2.0 617.0 ... 3.770249 617.7 97.6 2328.882761 0.000054
... ... ... ... ... ... ... ... ... ... ... ...
68952 1329.26058 617.0 2.000000 3.0 617.0 ... 1.200000 617.9 96.4 741.480000 2.526091
68953 1329.29780 617.0 2.000000 3.0 617.0 ... 1.200000 617.9 96.4 741.480000 2.526098
68954 1329.29836 617.0 2.000000 3.0 617.0 ... 1.200000 617.9 96.4 741.480000 2.526099
68955 1329.30060 617.0 2.000000 3.0 617.0 ... 1.200000 617.9 96.4 741.480000 2.526099
68956 1329.31193 617.0 2.000000 3.0 617.0 ... 1.200000 617.9 96.4 741.480000 2.526101
[68957 rows x 15 columns]
整车试验常用操作
行驶阻力系数
根据车辆总质量和车型，按照GB/T 18386-2017附录A中表格估算车辆行驶阻力系数：
f = xy.vehicle_test.driving_resistance(13500, 'coach')
>>>print(f)
2
0.154 x + 5.95 x + 751.3
>>>print(type(f))
<class 'numpy.poly1d'>
>>>print(f.c)
[1.5400e-01 5.9500e+00 7.5135e+02]
>>>print(f(60))
1662.75
稳态回转
稳态回转试验数据处理，主要参考：《GB/T 6323-2014 汽车操纵稳定性试验方法》10.4：
# 读取数据，第三个工作表中
df = processor.read('steady_static_circular_test.xlsx', 2)
# 移除数据中的无效值行
df = xy.data.remove_nan(df)
# 进行稳态回转的数据处理，df、轴距、各数据所在列名称
res = xy.vehicle_test.steady_static_circular_test(
df, 4.050, 'Time [s]', 'AccY_Comp [m/s/s]',
'V_X [km/h]', 'YAWRATE [deg/s]', 'Roll [deg]'
)
# 打印结果字典
for key, value in res.items():
print('{}:{}'.format(key, value))
>>>df: t a_y v_x ... r ratio_r d_sideslip_angle
0 3.76 -0.538704 8.191092 ... 13.898634 1.010649 0.177798
1 3.77 -0.517155 8.227018 ... 14.021179 1.019560 0.323719
2 3.78 -0.479446 8.262943 ... 13.699623 0.996178 -0.064736
3 3.79 -0.479446 8.298869 ... 13.858317 1.007718 0.129226
4 3.80 -0.506381 8.334795 ... 13.898283 1.010624 0.177376
... ... ... ... ... ... ...
1222 16.00 -4.374273 29.818448 ... 15.399734 1.119803 1.805226
1223 16.01 -4.530497 29.818448 ... 15.338217 1.115330 1.744792
1224 16.02 -4.568206 29.854374 ... 15.201510 1.105389 1.608740
1225 16.03 -4.315016 29.854374 ... 15.288837 1.111739 1.695929
1226 16.04 -4.234210 29.854374 ... 15.370341 1.117666 1.776411
[1227 rows x 9 columns]
charts:(<pygal.graph.xy.XY object at 0x000001B1736D8C10>, <pygal.graph.xy.XY object at 0x000001B1736D8BB0>, <pygal.graph.xy.XY object at 0x000001B172F35700>)
curves:(poly1d([ 3.88131566e-04, 2.98285198e-03, -1.73719579e-02, 1.00000000e+00]), poly1d([ 0.00586082, 0.03784224, -0.29914155, 0. ]), poly1d([-0.01398442, -0.87342555, -0.38406029]))
results:('left', 13.752182739265344, array([ 6.80479099, -2.50024479]), 0.35138275155465365, 0.8454567089306112)
# 打印试验结果
results = '方向：{}；\n初始圆半径：{} m；\n中性转向点侧向加速度：{} m/s^2；\n' \
'不足转向度：{} (°)/(m/s^2)；\n车身侧倾度：{} (°)/(m/s^2)。'.format(*res['results'])
>>>print(results)
方向：left；
初始圆半径：13.752182739265344 m；
中性转向点侧向加速度：[ 6.80479099 -2.50024479] m/s^2；
不足转向度：0.35138275155465365 (°)/(m/s^2)；
车身侧倾度：0.8454567089306112 (°)/(m/s^2)。
# 绘制试验结果曲线，res['charts']中依次为转向半径比特性图、前后轴侧偏角特性图、车厢侧倾角特性图
res['charts'][0].render_in_browser() # 渲染到默认浏览器
res['charts'][0].render_to_png('re.png') # 保存为png图片
res['charts'][0].render_to_file('re.svg') # 保存为svg图片
注意：保存为png格式的图片需要额外的程序支持，windows平台请下载安装gtk2-runtime-2.24.10-2012-10-10-ash.exe。
低速转向回正
低速转向回正试验数据处理，主要参考：《GB/T 6323-2014 汽车操纵稳定性试验方法》8.4：
# 读取数据，第三个工作表中
df = processor.read('returnability_test.xlsx', 2)
# 移除数据中的无效值行
df = xy.data.remove_nan(df)
# 进行低速转向回正的数据处理，df、松方向盘时刻、各数据所在列名称
res = xy.vehicle_test.returnability_test(df, 143.3, 'Time [s]', 'YAWRATE [deg/s]', 'ANGLE [deg]')
# 打印结果字典，图片分别为横摆角速度时域曲线、转向盘转角时域曲线，
# 结果分别为残余横摆角速度（正负表示方向）、横摆角速度总方差
for key, value in res.items():
print('{}:{}'.format(key, value))
>>>charts:(<pygal.graph.xy.XY object at 0x000001B1737AAF70>, <pygal.graph.xy.XY object at 0x000001B173756430>)
results:(-2.211031768936606, 0.8667891047150287)
转向轻便性
转向轻便性试验数据处理，主要参考：《GB/T 6323-2014 汽车操纵稳定性试验方法》9.4：
# 读取数据，第一个工作表中
df = processor.read('steering_efforts_test.xlsx', 0)
# 进行转向轻便性的数据处理，df、转向盘直径（m）、各数据所在列名称
res = xy.vehicle_test.steering_efforts_test(df, 0.48, 't', 'v_x', 'angle', 'torque')
# 打印结果字典，图片为转向盘力矩转角曲线，
# 结果分别为绕双纽线时的平均车速、左转转向盘最大转角、右转转向盘最大转角、转向盘最大作用力矩、转向盘最大作用力、绕双纽线路径一周的转向盘作用功、转向盘平均摩擦力矩、转向盘平均摩擦力
for key, value in res.items():
print('{}:{}'.format(key, value))
>>>results:(10.922474511888389, -645.404377133912, 661.401071702456, 6.60583377038265, 27.524307376594376, 94.27873879648446, 2.06678577735684, 8.611607405653501)
chart:<pygal.graph.xy.XY object at 0x000002AB49C38520>
注意：df数据需要自行剪切所需数据段后再传入函数，演示所用文件为经过剪切预处理后保存的数据，非原始数据。
双纽线摆桩
绘制双纽线摆桩图，主要参考：《GB/T 6323-2014 汽车操纵稳定性试验方法》9.3.1：
# 计算双纽线，第一个参数车辆最小转弯半径（左右转中取大的），第二个为车辆宽度
res = xy.vehicle_test.draw_lemniscate(15.94/2, width=2.2)
# 打印结果字典，图片为双纽线摆桩示意图，
# 结果分别为16个摆桩点坐标、双纽线最小曲率半径、x轴最大值、摆桩的偏移量、双纽线y轴最高点x坐标、双纽线y轴最高点y坐标
for key, value in res.items():
print('{}:{}'.format(key, value))
>>>chart:<pygal.graph.xy.XY object at 0x000002AB49C38AC0>
results:([(25.151000000000003, 0), (27.451, 0), (16.106007431235092, 8.148807725993692), (16.106007431235092, 10.448807725993692), (16.106007431235092, -10.448807725993692), (16.106007431235092, -8.148807725993692), (1.1500000000000001, 0), (-25.151000000000003, 0), (-27.451, 0), (-16.106007431235092, 8.148807725993692), (-16.106007431235092, 10.448807725993692), (-16.106007431235092, -10.448807725993692), (-16.106007431235092, -8.148807725993692), (-1.1500000000000001, 0), (0, -1.1500000000000001), (0, 1.1500000000000001)], 8.767000000000001, 26.301000000000002, 1.1500000000000001, 16.106007431235092, 9.298807725993692)
pdf常用操作
图片转pdf
推荐将需要转换合并的图片文件以001.jpg、002.jpg、003.jpg的形式顺序命名后放在同一文件夹下：
import glob
from xyw_utils.pdf import img_to_pdf, merge_pdfs
# 获取imgs文件夹下所有jpg后缀的文件名列表
imgs = glob.glob(r'imgs/*.jpg')
# 降序排列列表
imgs.sort()
# 将图片转为pdf并合并为一个文件
img_to_pdf(imgs, 'imgs.pdf')
拼接合并pdf文件
将多个pdf文件拼接合并到同一文件中：
# 在pdfs文件下创建五个pdf文件
for i in range(5):
img_to_pdf('imgs/00{}.jpg'.format(i + 1), 'pdfs/{}.pdf'.format(i + 1))
# 将刚才创建的五个文件合并为一个
merge_pdfs(['pdfs/1.pdf', 'pdfs/2.pdf', 'pdfs/3.pdf', 'pdfs/4.pdf', 'pdfs/5.pdf'], 'merge.pdf')
如果需要对pdf文件进行更多操作，可以使用pikepdf库，此处只是封装了其中两个小功能。
Release Notes
0.0.10
- vehicle_test模块新增按固定步长提取数据点的函数
- 新增充分发出的平均制动减速度的计算函数
0.0.9
- 修复了双纽线摆桩图生成函数中的标桩间隙错误
0.0.8
- 添加了说明文档《10分钟快速开始》
0.0.1
- 首次上传
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
xyw-utils-0.0.10.tar.gz
(2.1 MB
view hashes)
Built Distribution
Close
Hashes for xyw_utils-0.0.10-py3-none-any.whl
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 | e8f48dcde53f31b3dff2e044bfbcdd7aef65616ac8ea68bc40c76edc69a5b9d9 |
|
| MD5 | de2e1c9a938dae1bf5d256ea532aaacf |
|
| BLAKE2b-256 | 3b33e04962fa32a3c6125ffb2c447af478b6cf07c7be501cc36531c4d0ae1c3e |
