Jupyter Notebook中Python的Nyquist绘图函数报错求助

在Jupyter Notebook中调用control.nyquist_plot函数绘制Nyquist图时，程序抛出ValueError: array of sample points is empty错误，仅能显示部分图像，即使运行他人可正常执行的相同代码也会出现该问题。

复现代码 #

import numpy as np
import matplotlib.pyplot as plt
import control

sys = control.tf([0, 0, 10], [1, 3, 2])

fig = plt.figure(figsize=(10, 5))
control.nyquist_plot(sys);

报错信息 #

ValueError                                Traceback (most recent call last)
Cell In[3], line 4
1 sys = control.tf([0, 0, 10], [1, 3, 2])
3 fig = plt.figure(figsize=(10, 5))
----> 4 control.nyquist_plot(sys);

File ~.conda\envs\fondamenti-di-automatica\lib\site-packages\control\freqplot.py:1024, in nyquist_plot(syslist, omega, plot, omega_limits, omega_num, label_freq, color, return_contour, warn_encirclements, warn_nyquist, **kwargs)
1022 x_scl = np.ma.masked_where(scale_mask, resp.real)
1023 y_scl = np.ma.masked_where(scale_mask, resp.imag)
-> 1024 plt.plot(
1025     x_scl * (1 + curve_offset), y_scl * (1 + curve_offset),
1026     primary_style[1], color=c, **kwargs)

1028 # Plot the primary curve (invisible) for setting arrows
1029 x, y = resp.real.copy(), resp.imag.copy()

File ~.conda\envs\fondamenti-di-automatica\lib\site-packages\matplotlib\pyplot.py:2785, in plot(scalex, scaley, data, *args, **kwargs)
2783 @_copy_docstring_and_deprecators(Axes.plot)
2784 def plot(*args, scalex=True, scaley=True, data=None, **kwargs):
-> 2785     return gca().plot(
2786         *args, scalex=scalex, scaley=scaley,
2787         **({"data": data} if data is not None else {}), **kwargs)

File ~.conda\envs\fondamenti-di-automatica\lib\site-packages\matplotlib\axes\_axes.py:1690, in Axes.plot(self, scalex, scaley, data, *args, **kwargs)
1688 lines = [*self._get_lines(*args, data=data, **kwargs)]
1689 for line in lines:
-> 1690     self.add_line(line)
1691 if scalex:
1692     self._request_autoscale_view("x")

File ~.conda\envs\fondamenti-di-automatica\lib\site-packages\matplotlib\axes\_base.py:2304, in _AxesBase.add_line(self, line)
2301 if line.get_clip_path() is None:
2302     line.set_clip_path(self.patch)
-> 2304 self._update_line_limits(line)
2305 if not line.get_label():
2306     line.set_label(f'_child{len(self._children)}')

File ~.conda\envs\fondamenti-di-automatica\lib\site-packages\matplotlib\axes\_base.py:2327, in _AxesBase._update_line_limits(self, line)
2323 def _update_line_limits(self, line):
2324     """
2325     Figures out the data limit of the given line, updating self.dataLim.
2326     """
-> 2327     path = line.get_path()
2328     if path.vertices.size == 0:
2329         return

File ~.conda\envs\fondamenti-di-automatica\lib\site-packages\matplotlib\lines.py:1029, in Line2D.get_path(self)
1027 """Return the `~matplotlib.path.Path` associated with this line."""
1028 if self._invalidy or self._invalidx:
-> 1029     self.recache()
1030 return self._path

File ~.conda\envs\fondamenti-di-automatica\lib\site-packages\matplotlib\lines.py:681, in Line2D.recache(self, always)
679     self._x_filled = self._x.copy()
680     indices = np.arange(len(x))
--> 681     self._x_filled[nanmask] = np.interp(
682         indices[nanmask], indices[~nanmask], self._x[~nanmask])
683 else:
684     self._x_filled = self._x

File <__array_function__ internals>:200, in interp(*args, **kwargs)

File ~.conda\envs\fondamenti-di-automatica\lib\site-packages\numpy\lib\function_base.py:1595, in interp(x, xp, fp, left, right, period)
1592     xp = np.concatenate((xp[-1:]-period, xp, xp[0:1]+period))
1593     fp = np.concatenate((fp[-1:], fp, fp[0:1]))
-> 1595 return interp_func(x, xp, fp, left, right)

 
ValueError: array of sample points is empty

解决方案 #

• 降级control库版本：该错误多由control库0.10+版本与matplotlib、numpy的兼容性问题导致，降级到0.9.x版本（如0.9.3）可解决，执行命令：

  pip install control==0.9.3
  

• 手动指定频率范围：若不想降级库，可手动生成频率点传入omega参数，避免自动生成逻辑出现空值问题，修改后的代码如下：

  import numpy as np
  import matplotlib.pyplot as plt
  import control
  
  sys = control.tf([0, 0, 10], [1, 3, 2])
  
  fig = plt.figure(figsize=(10, 5))
  # 手动生成1000个从10^-2到10^2的对数分布频率点
  omega = np.logspace(-2, 2, 1000)
  control.nyquist_plot(sys, omega=omega);
  

内容的提问来源于stack exchange，提问作者Roberto