Description

hgducharme

opened on Aug 21, 2019

Issue body actions

I have step responses that go negative and sometimes stay negative, and it breaks the settling time calculation and rise time calculation. I've been messing around with it and might open a pull request if this is something y'all would be interested in.

For example, here is my code, the step response plot, and the error

from control import tf, step_response, step_info
import matplotlib.pyplot as plt
import numpy as np

transfer_function = tf( [3.32, 0, -162.8], [1, 24.56, 186.5, 457.8, 116.2] )
time, response = step_response(transfer_function)

fig, ax = plt.subplots(1, 1)
ax.plot(time, response)
plt.show()

step_info(transfer_function)

step_info() gives the following error

Traceback (most recent call last):
  File "bug_test.py", line 14, in <module>
    step_info(transfer_function)
  File "/Users/hgducharme/Programming/pendulum/env/lib/python3.7/site-packages/control/timeresp.py", line 536, in step_info
    SettlingTime = T[i + 1]
IndexError: index 1000 is out of bounds for axis 0 with size 1000

When I run this code in MATLAB, I get the following result

Activity

1. murrayrm commented on Aug 21, 2019

   

   murrayrm

   on Aug 21, 2019

   Member

   More actions

   @joaoantoniocardoso Any interest in taking a look at this?

   Reacted by João Antônio Cardoso
2. joaoantoniocardoso commented on Aug 21, 2019

   

   joaoantoniocardoso

   on Aug 21, 2019

   Contributor

   More actions

   @murrayrm Yes, I think I can take a look at this tomorrow.

   Also I think this error is not something that happens only for negative step responses... Months ago I did a pull-request (#293 ) that fixes some cases with the same message. (That pull was closed because while fixes some cases, it breaks others)
3. murrayrm

   added this to the 0.9.x milestone on Dec 26, 2020
4. murrayrm

   added

   bug

   enhancement

   on Dec 26, 2020
5. juanodecc commented on Mar 9, 2021

   

   juanodecc

   on Mar 9, 2021

   Contributor

   More actions

   Hi, think I solved this problem and the problem of systems with non settling time, I mean when dcgain isinf or isnan, doing:

       # Steady state value
       InfValue = sys1.dcgain()
       if InfValue < 0.0:
           yout = -yout
           InfValue = np.abs(InfValue)
   
   
   
   
       # SettlingTime
       sup_margin = (1. + SettlingTimeThreshold) * np.abs(InfValue)
       inf_margin = (1. - SettlingTimeThreshold) * np.abs(InfValue)
       # find Steady State looking for the first point out of specified limits
       if np.isnan(InfValue) or np.isinf(InfValue):
           SettlingTime = np.NaN
       else:
           for i in reversed(range(time.size-1)):
               if((np.abs(yout[i]) <= inf_margin) | (np.abs(yout[i]) >= sup_margin)):
                   SettlingTime = T[i + 1]
                   break
   
6. joaoantoniocardoso commented on Mar 9, 2021

   

   joaoantoniocardoso

   on Mar 9, 2021

   Contributor

   More actions

   Hi @juanodecc, it seems like a good solution. Could you open a PR with that, adding also a test case with a negative system and another with a system with non-settling time? Thanks
7. juanodecc

   added a commit that references this issue on Mar 12, 2021

   Solve issue python-control#337, python-control#565 and python-control…

   f964170
8. murrayrm

   linked a pull request that will close this issueStep info improve jpp #567on Mar 13, 2021
9. bnavigator

   closed this as completedin #567on Mar 17, 2021

Issue actions