# freqplot.py - frequency domain plots for control systems # # Author: Richard M. Murray # Date: 24 May 09 # # This file contains some standard control system plots: Bode plots, # Nyquist plots and pole-zero diagrams. The code for Nichols charts # is in nichols.py. # # Copyright (c) 2010 by California Institute of Technology # All rights reserved. # # Redistribution and use in source and binary forms, with or without # modification, are permitted provided that the following conditions # are met: # # 1. Redistributions of source code must retain the above copyright # notice, this list of conditions and the following disclaimer. # # 2. Redistributions in binary form must reproduce the above copyright # notice, this list of conditions and the following disclaimer in the # documentation and/or other materials provided with the distribution. # # 3. Neither the name of the California Institute of Technology nor # the names of its contributors may be used to endorse or promote # products derived from this software without specific prior # written permission. # # THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS # "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT # LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS # FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL CALTECH # OR THE CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, # SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT # LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF # USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND # ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, # OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT # OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF # SUCH DAMAGE. # # $Id$ import math import matplotlib as mpl import matplotlib.pyplot as plt import numpy as np import warnings from .ctrlutil import unwrap from .bdalg import feedback from .margins import stability_margins from .exception import ControlMIMONotImplemented from .statesp import StateSpace from .xferfcn import TransferFunction from . import config __all__ = ['bode_plot', 'nyquist_plot', 'gangof4_plot', 'singular_values_plot', 'bode', 'nyquist', 'gangof4'] # Default values for module parameter variables _freqplot_defaults = { 'freqplot.feature_periphery_decades': 1, 'freqplot.number_of_samples': 1000, 'freqplot.dB': False, # Plot gain in dB 'freqplot.deg': True, # Plot phase in degrees 'freqplot.Hz': False, # Plot frequency in Hertz 'freqplot.grid': True, # Turn on grid for gain and phase 'freqplot.wrap_phase': False, # Wrap the phase plot at a given value # deprecations 'deprecated.bode.dB': 'freqplot.dB', 'deprecated.bode.deg': 'freqplot.deg', 'deprecated.bode.Hz': 'freqplot.Hz', 'deprecated.bode.grid': 'freqplot.grid', 'deprecated.bode.wrap_phase': 'freqplot.wrap_phase', } # # Main plotting functions # # This section of the code contains the functions for generating # frequency domain plots # # # Bode plot # def bode_plot(syslist, omega=None, plot=True, omega_limits=None, omega_num=None, margins=None, method='best', *args, **kwargs): """Bode plot for a system Plots a Bode plot for the system over a (optional) frequency range. Parameters ---------- syslist : linsys List of linear input/output systems (single system is OK) omega : array_like List of frequencies in rad/sec to be used for frequency response dB : bool If True, plot result in dB. Default is false. Hz : bool If True, plot frequency in Hz (omega must be provided in rad/sec). Default value (False) set by config.defaults['freqplot.Hz'] deg : bool If True, plot phase in degrees (else radians). Default value (True) config.defaults['freqplot.deg'] plot : bool If True (default), plot magnitude and phase omega_limits : array_like of two values Limits of the to generate frequency vector. If Hz=True the limits are in Hz otherwise in rad/s. omega_num : int Number of samples to plot. Defaults to config.defaults['freqplot.number_of_samples']. margins : bool If True, plot gain and phase margin. method : method to use in computing margins (see :func:`stability_margins`) *args : :func:`matplotlib.pyplot.plot` positional properties, optional Additional arguments for `matplotlib` plots (color, linestyle, etc) **kwargs : :func:`matplotlib.pyplot.plot` keyword properties, optional Additional keywords (passed to `matplotlib`) Returns ------- mag : ndarray (or list of ndarray if len(syslist) > 1)) magnitude phase : ndarray (or list of ndarray if len(syslist) > 1)) phase in radians omega : ndarray (or list of ndarray if len(syslist) > 1)) frequency in rad/sec Other Parameters ---------------- grid : bool If True, plot grid lines on gain and phase plots. Default is set by `config.defaults['freqplot.grid']`. initial_phase : float Set the reference phase to use for the lowest frequency. If set, the initial phase of the Bode plot will be set to the value closest to the value specified. Units are in either degrees or radians, depending on the `deg` parameter. Default is -180 if wrap_phase is False, 0 if wrap_phase is True. wrap_phase : bool or float If wrap_phase is `False`, then the phase will be unwrapped so that it is continuously increasing or decreasing. If wrap_phase is `True` the phase will be restricted to the range [-180, 180) (or [:math:`-\\pi`, :math:`\\pi`) radians). If `wrap_phase` is specified as a float, the phase will be offset by 360 degrees if it falls below the specified value. Default to `False`, set by config.defaults['freqplot.wrap_phase']. The default values for Bode plot configuration parameters can be reset using the `config.defaults` dictionary, with module name 'bode'. Notes ----- 1. Alternatively, you may use the lower-level methods :meth:`LTI.frequency_response` or ``sys(s)`` or ``sys(z)`` or to generate the frequency response for a single system. 2. If a discrete time model is given, the frequency response is plotted along the upper branch of the unit circle, using the mapping ``z = exp(1j * omega * dt)`` where `omega` ranges from 0 to `pi/dt` and `dt` is the discrete timebase. If timebase not specified (``dt=True``), `dt` is set to 1. Examples -------- >>> sys = ss("1. -2; 3. -4", "5.; 7", "6. 8", "9.") >>> mag, phase, omega = bode(sys) """ # Make a copy of the kwargs dictionary since we will modify it kwargs = dict(kwargs) # Check to see if legacy 'Plot' keyword was used if 'Plot' in kwargs: import warnings warnings.warn("'Plot' keyword is deprecated in bode_plot; use 'plot'", FutureWarning) # Map 'Plot' keyword to 'plot' keyword plot = kwargs.pop('Plot') # Get values for params (and pop from list to allow keyword use in plot) dB = config._get_param( 'freqplot', 'dB', kwargs, _freqplot_defaults, pop=True) deg = config._get_param( 'freqplot', 'deg', kwargs, _freqplot_defaults, pop=True) Hz = config._get_param( 'freqplot', 'Hz', kwargs, _freqplot_defaults, pop=True) grid = config._get_param( 'freqplot', 'grid', kwargs, _freqplot_defaults, pop=True) plot = config._get_param('freqplot', 'plot', plot, True) margins = config._get_param( 'freqplot', 'margins', margins, False) wrap_phase = config._get_param( 'freqplot', 'wrap_phase', kwargs, _freqplot_defaults, pop=True) initial_phase = config._get_param( 'freqplot', 'initial_phase', kwargs, None, pop=True) omega_num = config._get_param('freqplot', 'number_of_samples', omega_num) # If argument was a singleton, turn it into a tuple if not hasattr(syslist, '__iter__'): syslist = (syslist,) omega, omega_range_given = _determine_omega_vector( syslist, omega, omega_limits, omega_num) if plot: # Set up the axes with labels so that multiple calls to # bode_plot will superimpose the data. This was implicit # before matplotlib 2.1, but changed after that (See # https://github.com/matplotlib/matplotlib/issues/9024). # The code below should work on all cases. # Get the current figure if 'sisotool' in kwargs: fig = kwargs['fig'] ax_mag = fig.axes[0] ax_phase = fig.axes[2] sisotool = kwargs['sisotool'] del kwargs['fig'] del kwargs['sisotool'] else: fig = plt.gcf() ax_mag = None ax_phase = None sisotool = False # Get the current axes if they already exist for ax in fig.axes: if ax.get_label() == 'control-bode-magnitude': ax_mag = ax elif ax.get_label() == 'control-bode-phase': ax_phase = ax # If no axes present, create them from scratch if ax_mag is None or ax_phase is None: plt.clf() ax_mag = plt.subplot(211, label='control-bode-magnitude') ax_phase = plt.subplot( 212, label='control-bode-phase', sharex=ax_mag) mags, phases, omegas, nyquistfrqs = [], [], [], [] for sys in syslist: if not sys.issiso(): # TODO: Add MIMO bode plots. raise ControlMIMONotImplemented( "Bode is currently only implemented for SISO systems.") else: omega_sys = np.asarray(omega) if sys.isdtime(strict=True): nyquistfrq = math.pi / sys.dt if not omega_range_given: # limit up to and including nyquist frequency omega_sys = np.hstack(( omega_sys[omega_sys < nyquistfrq], nyquistfrq)) else: nyquistfrq = None mag, phase, omega_sys = sys.frequency_response(omega_sys) mag = np.atleast_1d(mag) phase = np.atleast_1d(phase) # # Post-process the phase to handle initial value and wrapping # if initial_phase is None: # Start phase in the range 0 to -360 w/ initial phase = -180 # If wrap_phase is true, use 0 instead (phase \in (-pi, pi]) initial_phase = -math.pi if wrap_phase is not True else 0 elif isinstance(initial_phase, (int, float)): # Allow the user to override the default calculation if deg: initial_phase = initial_phase/180. * math.pi else: raise ValueError("initial_phase must be a number.") # Shift the phase if needed if abs(phase[0] - initial_phase) > math.pi: phase -= 2*math.pi * \ round((phase[0] - initial_phase) / (2*math.pi)) # Phase wrapping if wrap_phase is False: phase = unwrap(phase) # unwrap the phase elif wrap_phase is True: pass # default calculation OK elif isinstance(wrap_phase, (int, float)): phase = unwrap(phase) # unwrap the phase first if deg: wrap_phase *= math.pi/180. # Shift the phase if it is below the wrap_phase phase += 2*math.pi * np.maximum( 0, np.ceil((wrap_phase - phase)/(2*math.pi))) else: raise ValueError("wrap_phase must be bool or float.") mags.append(mag) phases.append(phase) omegas.append(omega_sys) nyquistfrqs.append(nyquistfrq) # Get the dimensions of the current axis, which we will divide up # TODO: Not current implemented; just use subplot for now if plot: nyquistfrq_plot = None if Hz: omega_plot = omega_sys / (2. * math.pi) if nyquistfrq: nyquistfrq_plot = nyquistfrq / (2. * math.pi) else: omega_plot = omega_sys if nyquistfrq: nyquistfrq_plot = nyquistfrq phase_plot = phase * 180. / math.pi if deg else phase mag_plot = mag if nyquistfrq_plot: # append data for vertical nyquist freq indicator line. # if this extra nyquist lime is is plotted in a single plot # command then line order is preserved when # creating a legend eg. legend(('sys1', 'sys2')) omega_nyq_line = np.array( (np.nan, nyquistfrq_plot, nyquistfrq_plot)) omega_plot = np.hstack((omega_plot, omega_nyq_line)) mag_nyq_line = np.array(( np.nan, 0.7*min(mag_plot), 1.3*max(mag_plot))) mag_plot = np.hstack((mag_plot, mag_nyq_line)) phase_range = max(phase_plot) - min(phase_plot) phase_nyq_line = np.array( (np.nan, min(phase_plot) - 0.2 * phase_range, max(phase_plot) + 0.2 * phase_range)) phase_plot = np.hstack((phase_plot, phase_nyq_line)) # # Magnitude plot # if dB: ax_mag.semilogx(omega_plot, 20 * np.log10(mag_plot), *args, **kwargs) else: ax_mag.loglog(omega_plot, mag_plot, *args, **kwargs) # Add a grid to the plot + labeling ax_mag.grid(grid and not margins, which='both') ax_mag.set_ylabel("Magnitude (dB)" if dB else "Magnitude") # # Phase plot # # Plot the data ax_phase.semilogx(omega_plot, phase_plot, *args, **kwargs) # Show the phase and gain margins in the plot if margins: # Compute stability margins for the system margin = stability_margins(sys, method=method) gm, pm, Wcg, Wcp = (margin[i] for i in (0, 1, 3, 4)) # Figure out sign of the phase at the first gain crossing # (needed if phase_wrap is True) phase_at_cp = phases[0][(np.abs(omegas[0] - Wcp)).argmin()] if phase_at_cp >= 0.: phase_limit = 180. else: phase_limit = -180. if Hz: Wcg, Wcp = Wcg/(2*math.pi), Wcp/(2*math.pi) # Draw lines at gain and phase limits ax_mag.axhline(y=0 if dB else 1, color='k', linestyle=':', zorder=-20) ax_phase.axhline(y=phase_limit if deg else math.radians(phase_limit), color='k', linestyle=':', zorder=-20) mag_ylim = ax_mag.get_ylim() phase_ylim = ax_phase.get_ylim() # Annotate the phase margin (if it exists) if pm != float('inf') and Wcp != float('nan'): if dB: ax_mag.semilogx( [Wcp, Wcp], [0., -1e5], color='k', linestyle=':', zorder=-20) else: ax_mag.loglog( [Wcp, Wcp], [1., 1e-8], color='k', linestyle=':', zorder=-20) if deg: ax_phase.semilogx( [Wcp, Wcp], [1e5, phase_limit + pm], color='k', linestyle=':', zorder=-20) ax_phase.semilogx( [Wcp, Wcp], [phase_limit + pm, phase_limit], color='k', zorder=-20) else: ax_phase.semilogx( [Wcp, Wcp], [1e5, math.radians(phase_limit) + math.radians(pm)], color='k', linestyle=':', zorder=-20) ax_phase.semilogx( [Wcp, Wcp], [math.radians(phase_limit) + math.radians(pm), math.radians(phase_limit)], color='k', zorder=-20) # Annotate the gain margin (if it exists) if gm != float('inf') and Wcg != float('nan'): if dB: ax_mag.semilogx( [Wcg, Wcg], [-20.*np.log10(gm), -1e5], color='k', linestyle=':', zorder=-20) ax_mag.semilogx( [Wcg, Wcg], [0, -20*np.log10(gm)], color='k', zorder=-20) else: ax_mag.loglog( [Wcg, Wcg], [1./gm, 1e-8], color='k', linestyle=':', zorder=-20) ax_mag.loglog( [Wcg, Wcg], [1., 1./gm], color='k', zorder=-20) if deg: ax_phase.semilogx( [Wcg, Wcg], [0, phase_limit], color='k', linestyle=':', zorder=-20) else: ax_phase.semilogx( [Wcg, Wcg], [0, math.radians(phase_limit)], color='k', linestyle=':', zorder=-20) ax_mag.set_ylim(mag_ylim) ax_phase.set_ylim(phase_ylim) if sisotool: ax_mag.text( 0.04, 0.06, 'G.M.: %.2f %s\nFreq: %.2f %s' % (20*np.log10(gm) if dB else gm, 'dB ' if dB else '', Wcg, 'Hz' if Hz else 'rad/s'), horizontalalignment='left', verticalalignment='bottom', transform=ax_mag.transAxes, fontsize=8 if int(mpl.__version__[0]) == 1 else 6) ax_phase.text( 0.04, 0.06, 'P.M.: %.2f %s\nFreq: %.2f %s' % (pm if deg else math.radians(pm), 'deg' if deg else 'rad', Wcp, 'Hz' if Hz else 'rad/s'), horizontalalignment='left', verticalalignment='bottom', transform=ax_phase.transAxes, fontsize=8 if int(mpl.__version__[0]) == 1 else 6) else: plt.suptitle( "Gm = %.2f %s(at %.2f %s), " "Pm = %.2f %s (at %.2f %s)" % (20*np.log10(gm) if dB else gm, 'dB ' if dB else '', Wcg, 'Hz' if Hz else 'rad/s', pm if deg else math.radians(pm), 'deg' if deg else 'rad', Wcp, 'Hz' if Hz else 'rad/s')) # Add a grid to the plot + labeling ax_phase.set_ylabel("Phase (deg)" if deg else "Phase (rad)") def gen_zero_centered_series(val_min, val_max, period): v1 = np.ceil(val_min / period - 0.2) v2 = np.floor(val_max / period + 0.2) return np.arange(v1, v2 + 1) * period if deg: ylim = ax_phase.get_ylim() ax_phase.set_yticks(gen_zero_centered_series( ylim[0], ylim[1], 45.)) ax_phase.set_yticks(gen_zero_centered_series( ylim[0], ylim[1], 15.), minor=True) else: ylim = ax_phase.get_ylim() ax_phase.set_yticks(gen_zero_centered_series( ylim[0], ylim[1], math.pi / 4.)) ax_phase.set_yticks(gen_zero_centered_series( ylim[0], ylim[1], math.pi / 12.), minor=True) ax_phase.grid(grid and not margins, which='both') # ax_mag.grid(which='minor', alpha=0.3) # ax_mag.grid(which='major', alpha=0.9) # ax_phase.grid(which='minor', alpha=0.3) # ax_phase.grid(which='major', alpha=0.9) # Label the frequency axis ax_phase.set_xlabel("Frequency (Hz)" if Hz else "Frequency (rad/sec)") if len(syslist) == 1: return mags[0], phases[0], omegas[0] else: return mags, phases, omegas # # Nyquist plot # # Default values for module parameter variables _nyquist_defaults = { 'nyquist.mirror_style': '--', 'nyquist.arrows': 2, 'nyquist.arrow_size': 8, 'nyquist.indent_radius': 1e-1, 'nyquist.indent_direction': 'right', } def nyquist_plot(syslist, omega=None, plot=True, omega_limits=None, omega_num=None, label_freq=0, color=None, return_contour=False, warn_nyquist=True, *args, **kwargs): """Nyquist plot for a system Plots a Nyquist plot for the system over a (optional) frequency range. The curve is computed by evaluating the Nyqist segment along the positive imaginary axis, with a mirror image generated to reflect the negative imaginary axis. Poles on or near the imaginary axis are avoided using a small indentation. The portion of the Nyquist contour at infinity is not explicitly computed (since it maps to a constant value for any system with a proper transfer function). Parameters ---------- syslist : list of LTI List of linear input/output systems (single system is OK). Nyquist curves for each system are plotted on the same graph. plot : boolean If True, plot magnitude omega : array_like Set of frequencies to be evaluated, in rad/sec. omega_limits : array_like of two values Limits to the range of frequencies. Ignored if omega is provided, and auto-generated if omitted. omega_num : int Number of frequency samples to plot. Defaults to config.defaults['freqplot.number_of_samples']. color : string Used to specify the color of the line and arrowhead. mirror_style : string or False Linestyle for mirror image of the Nyquist curve. If `False` then omit completely. Default linestyle ('--') is determined by config.defaults['nyquist.mirror_style']. return_contour : bool If 'True', return the contour used to evaluate the Nyquist plot. label_freq : int Label every nth frequency on the plot. If not specified, no labels are generated. arrows : int or 1D/2D array of floats Specify the number of arrows to plot on the Nyquist curve. If an integer is passed. that number of equally spaced arrows will be plotted on each of the primary segment and the mirror image. If a 1D array is passed, it should consist of a sorted list of floats between 0 and 1, indicating the location along the curve to plot an arrow. If a 2D array is passed, the first row will be used to specify arrow locations for the primary curve and the second row will be used for the mirror image. arrow_size : float Arrowhead width and length (in display coordinates). Default value is 8 and can be set using config.defaults['nyquist.arrow_size']. arrow_style : matplotlib.patches.ArrowStyle Define style used for Nyquist curve arrows (overrides `arrow_size`). indent_radius : float Amount to indent the Nyquist contour around poles that are at or near the imaginary axis. indent_direction : str For poles on the imaginary axis, set the direction of indentation to be 'right' (default), 'left', or 'none'. warn_nyquist : bool, optional If set to 'False', turn off warnings about frequencies above Nyquist. *args : :func:`matplotlib.pyplot.plot` positional properties, optional Additional arguments for `matplotlib` plots (color, linestyle, etc) **kwargs : :func:`matplotlib.pyplot.plot` keyword properties, optional Additional keywords (passed to `matplotlib`) Returns ------- count : int (or list of int if len(syslist) > 1) Number of encirclements of the point -1 by the Nyquist curve. If multiple systems are given, an array of counts is returned. contour : ndarray (or list of ndarray if len(syslist) > 1)), optional The contour used to create the primary Nyquist curve segment. To obtain the Nyquist curve values, evaluate system(s) along contour. Notes ----- 1. If a discrete time model is given, the frequency response is computed along the upper branch of the unit circle, using the mapping ``z = exp(1j * omega * dt)`` where `omega` ranges from 0 to `pi/dt` and `dt` is the discrete timebase. If timebase not specified (``dt=True``), `dt` is set to 1. 2. If a continuous-time system contains poles on or near the imaginary axis, a small indentation will be used to avoid the pole. The radius of the indentation is given by `indent_radius` and it is taken to the right of stable poles and the left of unstable poles. If a pole is exactly on the imaginary axis, the `indent_direction` parameter can be used to set the direction of indentation. Setting `indent_direction` to `none` will turn off indentation. If `return_contour` is True, the exact contour used for evaluation is returned. Examples -------- >>> sys = ss([[1, -2], [3, -4]], [[5], [7]], [[6, 8]], [[9]]) >>> count = nyquist_plot(sys) """ # Check to see if legacy 'Plot' keyword was used if 'Plot' in kwargs: warnings.warn("'Plot' keyword is deprecated in nyquist_plot; " "use 'plot'", FutureWarning) # Map 'Plot' keyword to 'plot' keyword plot = kwargs.pop('Plot') # Check to see if legacy 'labelFreq' keyword was used if 'labelFreq' in kwargs: warnings.warn("'labelFreq' keyword is deprecated in nyquist_plot; " "use 'label_freq'", FutureWarning) # Map 'labelFreq' keyword to 'label_freq' keyword label_freq = kwargs.pop('labelFreq') # Check to see if legacy 'arrow_width' or 'arrow_length' were used if 'arrow_width' in kwargs or 'arrow_length' in kwargs: warnings.warn( "'arrow_width' and 'arrow_length' keywords are deprecated in " "nyquist_plot; use `arrow_size` instead", FutureWarning) kwargs['arrow_size'] = \ (kwargs.get('arrow_width', 0) + kwargs.get('arrow_length', 0)) / 2 kwargs.pop('arrow_width', False) kwargs.pop('arrow_length', False) # Get values for params (and pop from list to allow keyword use in plot) omega_num = config._get_param('freqplot', 'number_of_samples', omega_num) mirror_style = config._get_param( 'nyquist', 'mirror_style', kwargs, _nyquist_defaults, pop=True) arrows = config._get_param( 'nyquist', 'arrows', kwargs, _nyquist_defaults, pop=True) arrow_size = config._get_param( 'nyquist', 'arrow_size', kwargs, _nyquist_defaults, pop=True) arrow_style = config._get_param('nyquist', 'arrow_style', kwargs, None) indent_radius = config._get_param( 'nyquist', 'indent_radius', kwargs, _nyquist_defaults, pop=True) indent_direction = config._get_param( 'nyquist', 'indent_direction', kwargs, _nyquist_defaults, pop=True) # If argument was a singleton, turn it into a list if not hasattr(syslist, '__iter__'): syslist = (syslist,) omega, omega_range_given = _determine_omega_vector( syslist, omega, omega_limits, omega_num) if not omega_range_given: # Start contour at zero frequency omega[0] = 0. # Go through each system and keep track of the results counts, contours = [], [] for sys in syslist: if not sys.issiso(): # TODO: Add MIMO nyquist plots. raise ControlMIMONotImplemented( "Nyquist plot currently only supports SISO systems.") # Figure out the frequency range omega_sys = np.asarray(omega) # Determine the contour used to evaluate the Nyquist curve if sys.isdtime(strict=True): # Transform frequencies in for discrete-time systems nyquistfrq = math.pi / sys.dt if not omega_range_given: # limit up to and including nyquist frequency omega_sys = np.hstack(( omega_sys[omega_sys < nyquistfrq], nyquistfrq)) # Issue a warning if we are sampling above Nyquist if np.any(omega_sys * sys.dt > np.pi) and warn_nyquist: warnings.warn("evaluation above Nyquist frequency") # do indentations in s-plane where it is more convenient splane_contour = 1j * omega_sys # Bend the contour around any poles on/near the imaginary axis # TODO: smarter indent radius that depends on dcgain of system # and timebase of discrete system. if isinstance(sys, (StateSpace, TransferFunction)) \ and indent_direction != 'none': if sys.isctime(): splane_poles = sys.pole() else: # map z-plane poles to s-plane, ignoring any at the origin # because we don't need to indent for them zplane_poles = sys.pole() zplane_poles = zplane_poles[~np.isclose(abs(zplane_poles), 0.)] splane_poles = np.log(zplane_poles)/sys.dt if splane_contour[1].imag > indent_radius \ and np.any(np.isclose(abs(splane_poles), 0)) \ and not omega_range_given: # add some points for quarter circle around poles at origin splane_contour = np.concatenate( (1j * np.linspace(0., indent_radius, 50), splane_contour[1:])) for i, s in enumerate(splane_contour): # Find the nearest pole p = splane_poles[(np.abs(splane_poles - s)).argmin()] # See if we need to indent around it if abs(s - p) < indent_radius: if p.real < 0 or (np.isclose(p.real, 0) \ and indent_direction == 'right'): # Indent to the right splane_contour[i] += \ np.sqrt(indent_radius ** 2 - (s-p).imag ** 2) elif p.real > 0 or (np.isclose(p.real, 0) \ and indent_direction == 'left'): # Indent to the left splane_contour[i] -= \ np.sqrt(indent_radius ** 2 - (s-p).imag ** 2) else: ValueError("unknown value for indent_direction") # change contour to z-plane if necessary if sys.isctime(): contour = splane_contour else: contour = np.exp(splane_contour * sys.dt) # Compute the primary curve resp = sys(contour) # Compute CW encirclements of -1 by integrating the (unwrapped) angle phase = -unwrap(np.angle(resp + 1)) count = int(np.round(np.sum(np.diff(phase)) / np.pi, 0)) counts.append(count) contours.append(contour) if plot: # Parse the arrows keyword if isinstance(arrows, int): N = arrows # Space arrows out, starting midway along each "region" arrow_pos = np.linspace(0.5/N, 1 + 0.5/N, N, endpoint=False) elif isinstance(arrows, (list, np.ndarray)): arrow_pos = np.sort(np.atleast_1d(arrows)) elif not arrows: arrow_pos = [] else: raise ValueError("unknown or unsupported arrow location") # Set the arrow style if arrow_style is None: arrow_style = mpl.patches.ArrowStyle( 'simple', head_width=arrow_size, head_length=arrow_size) # Save the components of the response x, y = resp.real, resp.imag # Plot the primary curve p = plt.plot(x, y, '-', color=color, *args, **kwargs) c = p[0].get_color() ax = plt.gca() _add_arrows_to_line2D( ax, p[0], arrow_pos, arrowstyle=arrow_style, dir=1) # Plot the mirror image if mirror_style is not False: p = plt.plot(x, -y, mirror_style, color=c, *args, **kwargs) _add_arrows_to_line2D( ax, p[0], arrow_pos, arrowstyle=arrow_style, dir=-1) # Mark the -1 point plt.plot([-1], [0], 'r+') # Label the frequencies of the points if label_freq: ind = slice(None, None, label_freq) for xpt, ypt, omegapt in zip(x[ind], y[ind], omega_sys[ind]): # Convert to Hz f = omegapt / (2 * np.pi) # Factor out multiples of 1000 and limit the # result to the range [-8, 8]. pow1000 = max(min(get_pow1000(f), 8), -8) # Get the SI prefix. prefix = gen_prefix(pow1000) # Apply the text. (Use a space before the text to # prevent overlap with the data.) # # np.round() is used because 0.99... appears # instead of 1.0, and this would otherwise be # truncated to 0. plt.text(xpt, ypt, ' ' + str(int(np.round(f / 1000 ** pow1000, 0))) + ' ' + prefix + 'Hz') if plot: ax = plt.gca() ax.set_xlabel("Real axis") ax.set_ylabel("Imaginary axis") ax.grid(color="lightgray") # "Squeeze" the results if len(syslist) == 1: counts, contours = counts[0], contours[0] # Return counts and (optionally) the contour we used return (counts, contours) if return_contour else counts # Internal function to add arrows to a curve def _add_arrows_to_line2D( axes, line, arrow_locs=[0.2, 0.4, 0.6, 0.8], arrowstyle='-|>', arrowsize=1, dir=1, transform=None): """ Add arrows to a matplotlib.lines.Line2D at selected locations. Parameters: ----------- axes: Axes object as returned by axes command (or gca) line: Line2D object as returned by plot command arrow_locs: list of locations where to insert arrows, % of total length arrowstyle: style of the arrow arrowsize: size of the arrow transform: a matplotlib transform instance, default to data coordinates Returns: -------- arrows: list of arrows Based on https://stackoverflow.com/questions/26911898/ """ if not isinstance(line, mpl.lines.Line2D): raise ValueError("expected a matplotlib.lines.Line2D object") x, y = line.get_xdata(), line.get_ydata() arrow_kw = { "arrowstyle": arrowstyle, } color = line.get_color() use_multicolor_lines = isinstance(color, np.ndarray) if use_multicolor_lines: raise NotImplementedError("multicolor lines not supported") else: arrow_kw['color'] = color linewidth = line.get_linewidth() if isinstance(linewidth, np.ndarray): raise NotImplementedError("multiwidth lines not supported") else: arrow_kw['linewidth'] = linewidth if transform is None: transform = axes.transData # Compute the arc length along the curve s = np.cumsum(np.sqrt(np.diff(x) ** 2 + np.diff(y) ** 2)) arrows = [] for loc in arrow_locs: n = np.searchsorted(s, s[-1] * loc) # Figure out what direction to paint the arrow if dir == 1: arrow_tail = (x[n], y[n]) arrow_head = (np.mean(x[n:n + 2]), np.mean(y[n:n + 2])) elif dir == -1: # Orient the arrow in the other direction on the segment arrow_tail = (x[n + 1], y[n + 1]) arrow_head = (np.mean(x[n:n + 2]), np.mean(y[n:n + 2])) else: raise ValueError("unknown value for keyword 'dir'") p = mpl.patches.FancyArrowPatch( arrow_tail, arrow_head, transform=transform, lw=0, **arrow_kw) axes.add_patch(p) arrows.append(p) return arrows # # Gang of Four plot # # TODO: think about how (and whether) to handle lists of systems def gangof4_plot(P, C, omega=None, **kwargs): """Plot the "Gang of 4" transfer functions for a system Generates a 2x2 plot showing the "Gang of 4" sensitivity functions [T, PS; CS, S] Parameters ---------- P, C : LTI Linear input/output systems (process and control) omega : array Range of frequencies (list or bounds) in rad/sec **kwargs : :func:`matplotlib.pyplot.plot` keyword properties, optional Additional keywords (passed to `matplotlib`) Returns ------- None """ if not P.issiso() or not C.issiso(): # TODO: Add MIMO go4 plots. raise ControlMIMONotImplemented( "Gang of four is currently only implemented for SISO systems.") # Get the default parameter values dB = config._get_param( 'freqplot', 'dB', kwargs, _freqplot_defaults, pop=True) Hz = config._get_param( 'freqplot', 'Hz', kwargs, _freqplot_defaults, pop=True) grid = config._get_param( 'freqplot', 'grid', kwargs, _freqplot_defaults, pop=True) # Compute the senstivity functions L = P * C S = feedback(1, L) T = L * S # Select a default range if none is provided # TODO: This needs to be made more intelligent if omega is None: omega = _default_frequency_range((P, C, S)) # Set up the axes with labels so that multiple calls to # gangof4_plot will superimpose the data. See details in bode_plot. plot_axes = {'t': None, 's': None, 'ps': None, 'cs': None} for ax in plt.gcf().axes: label = ax.get_label() if label.startswith('control-gangof4-'): key = label[len('control-gangof4-'):] if key not in plot_axes: raise RuntimeError( "unknown gangof4 axis type '{}'".format(label)) plot_axes[key] = ax # if any of the axes are missing, start from scratch if any((ax is None for ax in plot_axes.values())): plt.clf() plot_axes = {'s': plt.subplot(221, label='control-gangof4-s'), 'ps': plt.subplot(222, label='control-gangof4-ps'), 'cs': plt.subplot(223, label='control-gangof4-cs'), 't': plt.subplot(224, label='control-gangof4-t')} # # Plot the four sensitivity functions # omega_plot = omega / (2. * math.pi) if Hz else omega # TODO: Need to add in the mag = 1 lines mag_tmp, phase_tmp, omega = S.frequency_response(omega) mag = np.squeeze(mag_tmp) if dB: plot_axes['s'].semilogx(omega_plot, 20 * np.log10(mag), **kwargs) else: plot_axes['s'].loglog(omega_plot, mag, **kwargs) plot_axes['s'].set_ylabel("$|S|$" + " (dB)" if dB else "") plot_axes['s'].tick_params(labelbottom=False) plot_axes['s'].grid(grid, which='both') mag_tmp, phase_tmp, omega = (P * S).frequency_response(omega) mag = np.squeeze(mag_tmp) if dB: plot_axes['ps'].semilogx(omega_plot, 20 * np.log10(mag), **kwargs) else: plot_axes['ps'].loglog(omega_plot, mag, **kwargs) plot_axes['ps'].tick_params(labelbottom=False) plot_axes['ps'].set_ylabel("$|PS|$" + " (dB)" if dB else "") plot_axes['ps'].grid(grid, which='both') mag_tmp, phase_tmp, omega = (C * S).frequency_response(omega) mag = np.squeeze(mag_tmp) if dB: plot_axes['cs'].semilogx(omega_plot, 20 * np.log10(mag), **kwargs) else: plot_axes['cs'].loglog(omega_plot, mag, **kwargs) plot_axes['cs'].set_xlabel( "Frequency (Hz)" if Hz else "Frequency (rad/sec)") plot_axes['cs'].set_ylabel("$|CS|$" + " (dB)" if dB else "") plot_axes['cs'].grid(grid, which='both') mag_tmp, phase_tmp, omega = T.frequency_response(omega) mag = np.squeeze(mag_tmp) if dB: plot_axes['t'].semilogx(omega_plot, 20 * np.log10(mag), **kwargs) else: plot_axes['t'].loglog(omega_plot, mag, **kwargs) plot_axes['t'].set_xlabel( "Frequency (Hz)" if Hz else "Frequency (rad/sec)") plot_axes['t'].set_ylabel("$|T|$" + " (dB)" if dB else "") plot_axes['t'].grid(grid, which='both') plt.tight_layout() # # Singular values plot # def singular_values_plot(syslist, omega=None, plot=True, omega_limits=None, omega_num=None, *args, **kwargs): """Singular value plot for a system Plots a Singular Value plot for the system over a (optional) frequency range. Parameters ---------- syslist : linsys List of linear systems (single system is OK). omega : array_like List of frequencies in rad/sec to be used for frequency response. plot : bool If True (default), generate the singular values plot. omega_limits : array_like of two values Limits of the frequency vector to generate. If Hz=True the limits are in Hz otherwise in rad/s. omega_num : int Number of samples to plot. Default value (1000) set by config.defaults['freqplot.number_of_samples']. dB : bool If True, plot result in dB. Default value (False) set by config.defaults['freqplot.dB']. Hz : bool If True, plot frequency in Hz (omega must be provided in rad/sec). Default value (False) set by config.defaults['freqplot.Hz'] Returns ------- sigma : ndarray (or list of ndarray if len(syslist) > 1)) singular values omega : ndarray (or list of ndarray if len(syslist) > 1)) frequency in rad/sec Other Parameters ---------------- grid : bool If True, plot grid lines on gain and phase plots. Default is set by `config.defaults['freqplot.grid']`. Examples -------- >>> import numpy as np >>> den = [75, 1] >>> sys = TransferFunction([[[87.8], [-86.4]], [[108.2], [-109.6]]], [[den, den], [den, den]]) >>> omega = np.logspace(-4, 1, 1000) >>> sigma, omega = singular_values_plot(sys, plot=True) >>> singular_values_plot(sys, 0.0, plot=False) (array([[197.20868123], [ 1.39141948]]), array([0.])) """ # Make a copy of the kwargs dictionary since we will modify it kwargs = dict(kwargs) # Get values for params (and pop from list to allow keyword use in plot) dB = config._get_param( 'freqplot', 'dB', kwargs, _freqplot_defaults, pop=True) Hz = config._get_param( 'freqplot', 'Hz', kwargs, _freqplot_defaults, pop=True) grid = config._get_param( 'freqplot', 'grid', kwargs, _freqplot_defaults, pop=True) plot = config._get_param( 'freqplot', 'plot', plot, True) omega_num = config._get_param('freqplot', 'number_of_samples', omega_num) # If argument was a singleton, turn it into a tuple if not hasattr(syslist, '__iter__'): syslist = (syslist,) omega, omega_range_given = _determine_omega_vector( syslist, omega, omega_limits, omega_num) omega = np.atleast_1d(omega) if plot: fig = plt.gcf() ax_sigma = None # Get the current axes if they already exist for ax in fig.axes: if ax.get_label() == 'control-sigma': ax_sigma = ax # If no axes present, create them from scratch if ax_sigma is None: plt.clf() ax_sigma = plt.subplot(111, label='control-sigma') # color cycle handled manually as all singular values # of the same systems are expected to be of the same color color_cycle = plt.rcParams['axes.prop_cycle'].by_key()['color'] color_offset = 0 if len(ax_sigma.lines) > 0: last_color = ax_sigma.lines[-1].get_color() if last_color in color_cycle: color_offset = color_cycle.index(last_color) + 1 sigmas, omegas, nyquistfrqs = [], [], [] for idx_sys, sys in enumerate(syslist): omega_sys = np.asarray(omega) if sys.isdtime(strict=True): nyquistfrq = math.pi / sys.dt if not omega_range_given: # limit up to and including nyquist frequency omega_sys = np.hstack(( omega_sys[omega_sys < nyquistfrq], nyquistfrq)) omega_complex = np.exp(1j * omega_sys * sys.dt) else: nyquistfrq = None omega_complex = 1j*omega_sys fresp = sys(omega_complex, squeeze=False) fresp = fresp.transpose((2, 0, 1)) sigma = np.linalg.svd(fresp, compute_uv=False) sigmas.append(sigma.transpose()) # return shape is "channel first" omegas.append(omega_sys) nyquistfrqs.append(nyquistfrq) if plot: color = color_cycle[(idx_sys + color_offset) % len(color_cycle)] color = kwargs.pop('color', color) nyquistfrq_plot = None if Hz: omega_plot = omega_sys / (2. * math.pi) if nyquistfrq: nyquistfrq_plot = nyquistfrq / (2. * math.pi) else: omega_plot = omega_sys if nyquistfrq: nyquistfrq_plot = nyquistfrq sigma_plot = sigma if dB: ax_sigma.semilogx(omega_plot, 20 * np.log10(sigma_plot), color=color, *args, **kwargs) else: ax_sigma.loglog(omega_plot, sigma_plot, color=color, *args, **kwargs) if nyquistfrq_plot is not None: ax_sigma.axvline(x=nyquistfrq_plot, color=color) # Add a grid to the plot + labeling if plot: ax_sigma.grid(grid, which='both') ax_sigma.set_ylabel("Singular Values (dB)" if dB else "Singular Values") ax_sigma.set_xlabel("Frequency (Hz)" if Hz else "Frequency (rad/sec)") if len(syslist) == 1: return sigmas[0], omegas[0] else: return sigmas, omegas # # Utility functions # # This section of the code contains some utility functions for # generating frequency domain plots # # Determine the frequency range to be used def _determine_omega_vector(syslist, omega_in, omega_limits, omega_num): """Determine the frequency range for a frequency-domain plot according to a standard logic. If omega_in and omega_limits are both None, then omega_out is computed on omega_num points according to a default logic defined by _default_frequency_range and tailored for the list of systems syslist, and omega_range_given is set to False. If omega_in is None but omega_limits is an array-like of 2 elements, then omega_out is computed with the function np.logspace on omega_num points within the interval [min, max] = [omega_limits[0], omega_limits[1]], and omega_range_given is set to True. If omega_in is not None, then omega_out is set to omega_in, and omega_range_given is set to True Parameters ---------- syslist : list of LTI List of linear input/output systems (single system is OK) omega_in : 1D array_like or None Frequency range specified by the user omega_limits : 1D array_like or None Frequency limits specified by the user omega_num : int Number of points to be used for the frequency range (if the frequency range is not user-specified) Returns ------- omega_out : 1D array Frequency range to be used omega_range_given : bool True if the frequency range was specified by the user, either through omega_in or through omega_limits. False if both omega_in and omega_limits are None. """ omega_range_given = True if omega_in is None: if omega_limits is None: omega_range_given = False # Select a default range if none is provided omega_out = _default_frequency_range(syslist, number_of_samples=omega_num) else: omega_limits = np.asarray(omega_limits) if len(omega_limits) != 2: raise ValueError("len(omega_limits) must be 2") omega_out = np.logspace(np.log10(omega_limits[0]), np.log10(omega_limits[1]), num=omega_num, endpoint=True) else: omega_out = np.copy(omega_in) return omega_out, omega_range_given # Compute reasonable defaults for axes def _default_frequency_range(syslist, Hz=None, number_of_samples=None, feature_periphery_decades=None): """Compute a default frequency range for frequency domain plots. This code looks at the poles and zeros of all of the systems that we are plotting and sets the frequency range to be one decade above and below the min and max feature frequencies, rounded to the nearest integer. If no features are found, it returns logspace(-1, 1) Parameters ---------- syslist : list of LTI List of linear input/output systems (single system is OK) Hz : bool If True, the limits (first and last value) of the frequencies are set to full decades in Hz so it fits plotting with logarithmic scale in Hz otherwise in rad/s. Omega is always returned in rad/sec. number_of_samples : int, optional Number of samples to generate. The default value is read from ``config.defaults['freqplot.number_of_samples']. If None, then the default from `numpy.logspace` is used. feature_periphery_decades : float, optional Defines how many decades shall be included in the frequency range on both sides of features (poles, zeros). The default value is read from ``config.defaults['freqplot.feature_periphery_decades']``. Returns ------- omega : array Range of frequencies in rad/sec Examples -------- >>> from matlab import ss >>> sys = ss("1. -2; 3. -4", "5.; 7", "6. 8", "9.") >>> omega = _default_frequency_range(sys) """ # Set default values for options number_of_samples = config._get_param( 'freqplot', 'number_of_samples', number_of_samples) feature_periphery_decades = config._get_param( 'freqplot', 'feature_periphery_decades', feature_periphery_decades, 1) # Find the list of all poles and zeros in the systems features = np.array(()) freq_interesting = [] # detect if single sys passed by checking if it is sequence-like if not hasattr(syslist, '__iter__'): syslist = (syslist,) for sys in syslist: try: # Add new features to the list if sys.isctime(): features_ = np.concatenate((np.abs(sys.pole()), np.abs(sys.zero()))) # Get rid of poles and zeros at the origin toreplace = features_ == 0.0 if np.any(toreplace): features_ = features_[~toreplace] elif sys.isdtime(strict=True): fn = math.pi * 1. / sys.dt # TODO: What distance to the Nyquist frequency is appropriate? freq_interesting.append(fn * 0.9) features_ = np.concatenate((sys.pole(), sys.zero())) # Get rid of poles and zeros on the real axis (imag==0) # * origin and real < 0 # * at 1.: would result in omega=0. (logaritmic plot!) toreplace = (features_.imag == 0.0) & ( (features_.real <= 0.) | (np.abs(features_.real - 1.0) < 1.e-10)) if np.any(toreplace): features_ = features_[~toreplace] # TODO: improve features_ = np.abs(np.log(features_) / (1.j * sys.dt)) else: # TODO raise NotImplementedError( "type of system in not implemented now") features = np.concatenate((features, features_)) except NotImplementedError: pass # Make sure there is at least one point in the range if features.shape[0] == 0: features = np.array([1.]) if Hz: features /= 2. * math.pi features = np.log10(features) lsp_min = np.floor(np.min(features) - feature_periphery_decades) lsp_max = np.ceil(np.max(features) + feature_periphery_decades) lsp_min += np.log10(2. * math.pi) lsp_max += np.log10(2. * math.pi) else: features = np.log10(features) lsp_min = np.floor(np.min(features) - feature_periphery_decades) lsp_max = np.ceil(np.max(features) + feature_periphery_decades) if freq_interesting: lsp_min = min(lsp_min, np.log10(min(freq_interesting))) lsp_max = max(lsp_max, np.log10(max(freq_interesting))) # TODO: Add a check in discrete case to make sure we don't get aliasing # (Attention: there is a list of system but only one omega vector) # Set the range to be an order of magnitude beyond any features if number_of_samples: omega = np.logspace( lsp_min, lsp_max, num=number_of_samples, endpoint=True) else: omega = np.logspace(lsp_min, lsp_max, endpoint=True) return omega # # Utility functions to create nice looking labels (KLD 5/23/11) # def get_pow1000(num): """Determine exponent for which significand of a number is within the range [1, 1000). """ # Based on algorithm from http://www.mail-archive.com/ # matplotlib-users@lists.sourceforge.net/msg14433.html, accessed 2010/11/7 # by Jason Heeris 2009/11/18 from decimal import Decimal from math import floor dnum = Decimal(str(num)) if dnum == 0: return 0 elif dnum < 0: dnum = -dnum return int(floor(dnum.log10() / 3)) def gen_prefix(pow1000): """Return the SI prefix for a power of 1000. """ # Prefixes according to Table 5 of [BIPM 2006] (excluding hecto, # deca, deci, and centi). if pow1000 < -8 or pow1000 > 8: raise ValueError( "Value is out of the range covered by the SI prefixes.") return ['Y', # yotta (10^24) 'Z', # zetta (10^21) 'E', # exa (10^18) 'P', # peta (10^15) 'T', # tera (10^12) 'G', # giga (10^9) 'M', # mega (10^6) 'k', # kilo (10^3) '', # (10^0) 'm', # milli (10^-3) r'$\mu$', # micro (10^-6) 'n', # nano (10^-9) 'p', # pico (10^-12) 'f', # femto (10^-15) 'a', # atto (10^-18) 'z', # zepto (10^-21) 'y'][8 - pow1000] # yocto (10^-24) def find_nearest_omega(omega_list, omega): omega_list = np.asarray(omega_list) return omega_list[(np.abs(omega_list - omega)).argmin()] # Function aliases bode = bode_plot nyquist = nyquist_plot gangof4 = gangof4_plot