def test_from_zpk(self):
# 1st order low-pass filter: H(s) = 0.3 / (z - 0.2),
system_ZPK = dlti([],[0.2],0.3)
system_TF = dlti(0.3, [1, -0.2])
w = [0.1, 1, 10, 100]
w1, H1 = dfreqresp(system_ZPK, w=w)
w2, H2 = dfreqresp(system_TF, w=w)
assert_almost_equal(H1, H2)
def test_from_zpk(self):
# 1st order low-pass filter: H(s) = 0.3 / (z - 0.2),
system_ZPK = dlti([],[0.2],0.3)
system_TF = dlti(0.3, [1, -0.2])
w = [0.1, 1, 10, 100]
w1, H1 = dfreqresp(system_ZPK, w=w)
w2, H2 = dfreqresp(system_TF, w=w)
assert_almost_equal(H1, H2)
def test_pole_one(self):
# Test that freqresp() doesn't fail on a system with a pole at 0.
# integrator, pole at zero: H(s) = 1 / s
system = TransferFunction([1], [1, -1], dt=0.1)
with warnings.catch_warnings():
warnings.simplefilter("ignore", RuntimeWarning)
w, H = dfreqresp(system, n=2)
assert_equal(w[0], 0.) # a fail would give not-a-number
def test_from_state_space(self):
# H(z) = 2 / z^3 - 0.5 * z^2
system_TF = dlti([2], [1, -0.5, 0, 0])
A = np.array([[0.5, 0, 0], [1, 0, 0], [0, 1, 0]])
B = np.array([[1, 0, 0]]).T
C = np.array([[0, 0, 2]])
D = 0
system_SS = dlti(A, B, C, D)
w = 10.0**np.arange(-3, 0, .5)
with suppress_warnings() as sup:
sup.filter(BadCoefficients)
w1, H1 = dfreqresp(system_TF, w=w)
w2, H2 = dfreqresp(system_SS, w=w)
assert_almost_equal(H1, H2)
def test_pole_one(self):
# Test that freqresp() doesn't fail on a system with a pole at 0.
# integrator, pole at zero: H(s) = 1 / s
system = TransferFunction([1], [1, -1], dt=0.1)
with warnings.catch_warnings():
warnings.simplefilter("ignore", RuntimeWarning)
w, H = dfreqresp(system, n=2)
assert_equal(w[0], 0.) # a fail would give not-a-number
def test_freq_range(self):
# Test that freqresp() finds a reasonable frequency range.
# 1st order low-pass filter: H(z) = 1 / (z - 0.2),
# Expected range is from 0.01 to 10.
system = TransferFunction(1, [1, -0.2], dt=0.1)
n = 10
expected_w = np.linspace(0, np.pi, 10, endpoint=False)
w, H = dfreqresp(system, n=n)
assert_almost_equal(w, expected_w)
def test_freq_range(self):
# Test that freqresp() finds a reasonable frequency range.
# 1st order low-pass filter: H(z) = 1 / (z - 0.2),
# Expected range is from 0.01 to 10.
system = TransferFunction(1, [1, -0.2], dt=0.1)
n = 10
expected_w = np.linspace(0, np.pi, 10, endpoint=False)
w, H = dfreqresp(system, n=n)
assert_almost_equal(w, expected_w)
def test_pole_one(self):
# Test that freqresp() doesn't fail on a system with a pole at 0.
# integrator, pole at zero: H(s) = 1 / s
system = TransferFunction([1], [1, -1], dt=0.1)
with suppress_warnings() as sup:
sup.filter(RuntimeWarning, message="divide by zero")
sup.filter(RuntimeWarning, message="invalid value encountered")
w, H = dfreqresp(system, n=2)
assert_equal(w[0], 0.) # a fail would give not-a-number
def test_pole_one(self):
# Test that freqresp() doesn't fail on a system with a pole at 0.
# integrator, pole at zero: H(s) = 1 / s
system = TransferFunction([1], [1, -1], dt=0.1)
with suppress_warnings() as sup:
sup.filter(RuntimeWarning, message="divide by zero")
sup.filter(RuntimeWarning, message="invalid value encountered")
w, H = dfreqresp(system, n=2)
assert_equal(w[0], 0.) # a fail would give not-a-number
def test_from_state_space(self):
# H(z) = 2 / z^3 - 0.5 * z^2
system_TF = dlti([2], [1, -0.5, 0, 0])
A = np.array([[0.5, 0, 0],
[1, 0, 0],
[0, 1, 0]])
B = np.array([[1, 0, 0]]).T
C = np.array([[0, 0, 2]])
D = 0
system_SS = dlti(A, B, C, D)
w = 10.0**np.arange(-3,0,.5)
with warnings.catch_warnings():
warnings.simplefilter("ignore", BadCoefficients)
w1, H1 = dfreqresp(system_TF, w=w)
w2, H2 = dfreqresp(system_SS, w=w)
assert_almost_equal(H1, H2)
def test_manual(self):
# Test dfreqresp() real part calculation (manual sanity check).
# 1st order low-pass filter: H(z) = 1 / (z - 0.2),
system = TransferFunction(1, [1, -0.2], dt=0.1)
w = [0.1, 1, 10]
w, H = dfreqresp(system, w=w)
# test real
expected_re = [1.2383, 0.4130, -0.7553]
assert_almost_equal(H.real, expected_re, decimal=4)
# test imag
expected_im = [-0.1555, -1.0214, 0.3955]
assert_almost_equal(H.imag, expected_im, decimal=4)
def test_manual(self):
# Test dfreqresp() real part calculation (manual sanity check).
# 1st order low-pass filter: H(z) = 1 / (z - 0.2),
system = TransferFunction(1, [1, -0.2], dt=0.1)
w = [0.1, 1, 10]
w, H = dfreqresp(system, w=w)
# test real
expected_re = [1.2383, 0.4130, -0.7553]
assert_almost_equal(H.real, expected_re, decimal=4)
# test imag
expected_im = [-0.1555, -1.0214, 0.3955]
assert_almost_equal(H.imag, expected_im, decimal=4)
def test_auto(self):
# Test dfreqresp() real part calculation.
# 1st order low-pass filter: H(z) = 1 / (z - 0.2),
system = TransferFunction(1, [1, -0.2], dt=0.1)
w = [0.1, 1, 10, 100]
w, H = dfreqresp(system, w=w)
jw = np.exp(w * 1j)
y = np.polyval(system.num, jw) / np.polyval(system.den, jw)
# test real
expected_re = y.real
assert_almost_equal(H.real, expected_re)
# test imag
expected_im = y.imag
assert_almost_equal(H.imag, expected_im)
def test_auto(self):
# Test dfreqresp() real part calculation.
# 1st order low-pass filter: H(z) = 1 / (z - 0.2),
system = TransferFunction(1, [1, -0.2], dt=0.1)
w = [0.1, 1, 10, 100]
w, H = dfreqresp(system, w=w)
jw = np.exp(w * 1j)
y = np.polyval(system.num, jw) / np.polyval(system.den, jw)
# test real
expected_re = y.real
assert_almost_equal(H.real, expected_re)
# test imag
expected_im = y.imag
assert_almost_equal(H.imag, expected_im)
t = t1 * t2 * t3
tf = algebra.conseguir_tf(t, s)
k = 1e3
fs = 20 * k
tf2 = signal.dlti(*signal.bilinear(tf.num, tf.den, fs))
w_range = linspace(0, fs, 100000) * 2 * pi
w, h = signal.freqresp(tf, w_range)
w2, h2 = signal.dfreqresp(tf2, w_range / fs)
f = w / 2 / pi
f2 = w2 / 2 / pi
CombinedPlot() \
.setTitle("Legendre") \
.setXTitle("Frecuencia (hz)") \
.setYTitle("Amplitud (Db)") \
.addSignalPlot(
signal=Senial.Senial(
f, 20 * log10(abs(h))
),
color="red",
name="Analógica"
) \
# Generating the Nyquist plot of a transfer function from scipy import signal import matplotlib.pyplot as plt # Transfer function: H(z) = 1 / (z^2 + 2z + 3) sys = signal.TransferFunction([1], [1, 2, 3], dt=0.05) w, H = signal.dfreqresp(sys) plt.figure() plt.plot(H.real, H.imag, "b") plt.plot(H.real, -H.imag, "r") plt.show()
# make the panes transparent
ax.xaxis.set_pane_color((1.0, 1.0, 1.0, 0.0))
ax.yaxis.set_pane_color((1.0, 1.0, 1.0, 0.0))
ax.zaxis.set_pane_color((1.0, 1.0, 1.0, 0.0))
plt.xlabel("Im(z)")
plt.ylabel("Re(z)")
plt.savefig("../fig/z_plane_%d.png" % indice)
# Réponse impulsionnelle
plt.figure()
t, yout = sig.dimpulse(tf_temp, n=30)
plt.stem(t, yout[0])
np.savetxt("../csv/ri_%d.txt" % indice, yout[0], delimiter=',')
# Réponse frequentielle
w, h = sig.dfreqresp(tf_temp, n=200)
amp = abs(h)
angles = np.unwrap(np.angle(h))
M = np.matrix([w, amp, angles]).T
plt.figure()
plt.plot(w, amp)
plt.figure()
plt.plot(w, angles)
np.savetxt("../csv/rf_%d.txt" % indice, M, delimiter=',')
#reponse à une sinusoide
output = sig.lfilter(tf_temp.num, tf_temp.den, input)
M = np.array([n_vect, input, output]).T
np.savetxt("../csv/r_sine_%d.txt" % indice, M, delimiter=',')
plt.show()
plot_dbode_ML(b_coef_traces[:, -1::-1],
f_coef_traces[:, -1::-1],
b_true.flatten(),
f_true.flatten(),
B_ML.flatten(),
F_ML.flatten(),
B_ML2.flatten(),
F_ML2.flatten(),
Ts,
w_plot,
save=True)
import scipy.signal as signal
w, _ = signal.dfreqresp((b_true.flatten(), f_true.flatten(), Ts))
# plot_d_nyquist(b_coef_traces[:,-1::-1],f_coef_traces[:,-1::-1],b_true.flatten(),f_true.flatten(),B_unit.flatten(),F_unit.flatten(),Ts,w, no_plot=300,xlims=[-2.75, 1.5], ylims=[-2,2], save='figures/oe_nyquist.png')
plot_d_nyquist(b_coef_traces[:, -1::-1],
f_coef_traces[:, -1::-1],
b_true.flatten(),
f_true.flatten(),
B_ML2.flatten(),
F_ML2.flatten(),
Ts,
w,
no_plot=300,
xlims=[-2.75, 1.5],
ylims=[-2, 2],
save='figures/oe_nyquist.png')
# b_mean = np.mean(b_coef_traces,axis=0)
Hs = (s_ * R_ * C_) / (s_**2 * L_ * C_ + s_ * R_ * C_ + 1)
print("H(s)=", Hs)
Hz = Hs.copy().subs(s_, 2 / T_ * (z_ - 1) / (z_ + 1))
print("H(z)=", Hz.cancel().collect(z_))
systemD = signal.cont2discrete((system.num, system.den),
1 / Fs,
method='bilinear')
systemD = (systemD[0][0], systemD[1], systemD[2])
plt.figure(figsize=(8, 3))
plt.subplot(121)
w = 2 * np.pi * np.linspace(100, 2e4, 50000) / Fs
_, gd = signal.group_delay(systemD[:2], w=w) # calculate group delay
_, H = signal.dfreqresp(systemD, w=w) # calculate frequency response
# calculate phase and group delay at the specific frequencies
indf0 = np.argmin(abs(w - 2 * np.pi * f0 / Fs))
indf1 = np.argmin(abs(w - 2 * np.pi * f1 / Fs))
tau_gf0 = gd[indf0] * 1000 / Fs
tau_gf1 = gd[indf1] * 1000 / Fs
tau_phif0 = (-np.angle(H[indf0]) / w[indf0] / Fs * 1000)
tau_phif1 = (-np.angle(H[indf1]) / w[indf1] / Fs * 1000)
plt.plot(w / (2 * np.pi) * Fs, gd / Fs * 1000, 'navy')
plt.grid(True)
plt.xlabel('$f$ [Hz]')
plt.ylabel(r'$\tau_g$ [ms]')
plt.title("Group Delay")
plt.axhline(tau_gf0, color='r', ls='dashed')
def calcularPlotImpinvar(f0,
fs,
mode="butter",
n=4,
filename="out.png",
title="noTitle"):
w0 = 2 * pi * f0
print("f0 = ", f0)
if mode == "butter":
b, a = signal.butter(n, w0, 'low', analog=True)
elif mode == "cheby":
b, a = signal.cheby1(n, 1, w0, 'low', analog=True)
sys = signal.lti(b, a)
w_range = linspace(0, fs, 100000) * 2 * pi
w, h = signal.freqresp(sys, w_range) # signal.freqs(b, a, 100000)
w_m2 = -1
for i in range(len(w)):
if 20 * log10(abs(h[i])) <= -2 and w_m2 == -1:
w_m2 = w[i]
if mode == "butter":
b, a = signal.butter(n, w0 * (w0 / w_m2), 'low', analog=True)
elif mode == "cheby":
b, a = signal.cheby1(n, 1, w0 * (w0 / w_m2), 'low', analog=True)
sys = signal.lti(b, a)
sys_original = sys
w, h = signal.freqresp(sys, w_range)
f = w / 2 / pi
b2, a2 = impinvar_causal(b, a, fs=fs, tol=0.0001)
sys = signal.dlti(b2, a2)
w2, h2 = signal.dfreqresp(sys, w_range / fs)
factor = h[0] / h2[0]
sys = signal.dlti(b2 / factor, a2)
f2 = w2 / 2 / pi * fs
CombinedPlot() \
.setTitle(title) \
.setXTitle("Frecuencia (hz)") \
.setYTitle("Amplitud (Db)") \
.addSignalPlot(
signal=Senial.Senial(
f, 20 * log10(abs(h))
),
color="red",
name="Analógica"
).addSignalPlot(
signal=Senial.Senial(
f2, 20 * log10(abs(h2) * factor)
),
color="blue",
name="Digital método invariante al impulso"
).plot().save("output/" + filename)
""" Impulse Invariance"""
num = np.array([0.000000000001, 0.00931, 0.000000000001])
den = np.array([1, -1.8588, 0.8681])
H = sig.TransferFunction(num, den, dt=Te)
print(H)
"""Bilinear transform"""
c = wc / np.tan(wc * Te / 2)
num = np.array([10000, 20000, 10000])
den = np.array([(10000 - 200 * c * np.cos(3 * np.pi / 4) + c * c),
(20000 - 2 * c * c),
(10000 + 200 * c * np.cos(3 * np.pi / 4) + c * c)])
H2 = sig.TransferFunction(num, den, dt=Te)
# Affichage
w_vect = np.logspace(1, 3.49, 1000)
w, Hij = sig.dfreqresp(H, w=w_vect * Te)
wb, Hbj = sig.dfreqresp(H2, w=w_vect * Te)
wa, Haj = sig.freqresp(Ha, w=w_vect)
plt.semilogx(wa, 20 * np.log10(np.abs(Hij)))
plt.semilogx(wa, 20 * np.log10(np.abs(Hbj)))
plt.semilogx(wa, 20 * np.log10(np.abs(Haj)))
plt.figure()
plt.semilogx(wa, np.unwrap(np.angle(Hij)))
plt.semilogx(wa, np.unwrap(np.angle(Hbj)))
plt.semilogx(wa, np.angle(Haj))
M = np.matrix([
wa, 20 * np.log10(np.abs(Haj)), 20 * np.log10(np.abs(Hij)),
20 * np.log10(np.abs(Hbj))
resbrute = optimize.brute(cost_function,
rranges,
full_output=True,
finish=optimize.fmin)
print(resbrute[0])
x0 = resbrute[0]
res = minimize(cost_function, x0)
omega = res.x
print(omega)
num, den = extract_num_den(omega)
H = sig.TransferFunction(num, den, dt=Te)
w_vect = np.logspace(1, 3.49, 100)
wb, Hj = sig.dfreqresp(H, w=w_vect * Te)
t, rit = sig.dimpulse(H, n=200)
plt.figure()
plt.plot(np.ravel(rit))
plt.figure()
plt.semilogx(w_vect, 20 * np.log10(np.abs(Hj)))
RF = np.zeros((len(w_vect), 2))
RF[:, 0] = w_vect
RF[:, 1] = 20 * np.log10(np.abs(Hj))
#np.savetxt("../csv/iir_prony_rf.txt",RF,delimiter=',')
plt.show()
def plot_d_nyquist(num_samples,
den_samples,
num_true,
den_true,
num_ML,
den_ML,
Ts,
omega,
no_plot=300,
xlims=None,
ylims=None,
max_samples=1000,
save=False):
"""plot nyquist diagram from estimated discrete time system samples and true sys"""
no_samples = np.shape(num_samples)[0]
no_eval = min(no_samples, max_samples)
sel = np.random.choice(np.arange(no_samples), no_eval, False)
omega_res = max(np.shape(omega))
H_real_samples = np.zeros((omega_res, no_eval))
H_imag_samples = np.zeros((omega_res, no_eval))
count = 0
for s in sel:
den_sample = np.concatenate(([1.0], den_samples[s, :]), 0)
num_sample = num_samples[s, :]
w, H_samp = signal.dfreqresp((num_sample, den_sample, Ts), omega)
H_real_samples[:, count] = H_samp.real
H_imag_samples[:, count] = H_samp.imag
# w, mag_samples[:, count], phase_samples[:, count] = signal.dbode((num_sample, den_sample, Ts), omega)
count = count + 1
# calculate the true bode diagram
# plot the true bode diagram
w, H_true = signal.dfreqresp((num_true.flatten(), den_true.flatten(), Ts),
omega)
w, H_ML = signal.dfreqresp((num_ML.flatten(), den_ML.flatten(), Ts), omega)
# plot the samples
plt.subplot(1, 1, 1)
h2, = plt.plot(H_real_samples[:, 0],
H_imag_samples[:, 0],
color='green',
alpha=0.1,
label='hmc samples') # Bode magnitude plot
plt.plot(H_real_samples[:, 1:no_plot],
H_imag_samples[:, 1:no_plot],
color='green',
alpha=0.1) # Bode magnitude plot
# plt.plot(H_real_samples[:,:no_plot], -H_imag_samples[:,:no_plot], color='green', alpha=0.1) # Bode magnitude plot
h1, = plt.plot(H_true.real, H_true.imag, color='blue',
label='True system') # Bode magnitude plot
# plt.plot(H_true.real, -H_true.imag, color='blue') # Bode magnitude plot
h_ML, = plt.plot(H_ML.real,
H_ML.imag,
'--',
color='purple',
label='ML Estimate') # Bode magnitude plot
# plt.plot(H_ML.real, -H_ML.imag,'--', color='purple') # Bode magnitude plot
hm, = plt.plot(np.mean(H_real_samples, axis=1),
np.mean(H_imag_samples, axis=1),
'-.',
color='orange',
label='hmc mean') # Bode magnitude plot
# plt.plot(np.mean(H_real_samples,axis=1), -np.mean(H_imag_samples,axis=1), '-.', color='orange') # Bode magnitude plot
plt.xlabel('Real')
plt.ylabel('Imaginary')
if xlims is not None:
plt.xlim(xlims)
if ylims is not None:
plt.ylim(ylims)
plt.legend(handles=[h1, h2, hm, h_ML])
if save is not None:
plt.savefig(save, format='png')
plt.show()
def calcularPlotMatchedZ(f0, fs, mode="butter", n=4, filename="out.png", title="noTitle"):
w0 = 2 * pi * f0
if mode == "butter":
b, a = signal.butter(n, w0, 'low', analog=True)
elif mode == "cheby":
b, a = signal.cheby1(n, 1, w0, 'low', analog=True)
sys = signal.lti(b, a)
w_range = linspace(0, fs, 100000) * 2 * pi
w, h = signal.freqresp(sys, w_range) # signal.freqs(b, a, 100000)
w_m2 = -1
for i in range(len(w)):
if 20 * log10(abs(h[i])) <= -2 and w_m2 == -1:
w_m2 = w[i]
if mode == "butter":
b, a = signal.butter(n, w0 * (w0 / w_m2), 'low', analog=True)
elif mode == "cheby":
b, a = signal.cheby1(n, 1, w0 * (w0 / w_m2), 'low', analog=True)
sys = signal.lti(b, a)
w, h = signal.freqresp(sys, w_range)
f = w / 2 / pi
poles = sys.poles
zeros = sys.zeros
new_poles = exp(poles / fs)
new_zeros = exp(zeros / fs)
new_zeros = np.hstack([new_zeros, [-1] * n])
# print(new_poles, new_zeros)
sys = signal.dlti(new_zeros, new_poles, 1)
w2, h2 = signal.dfreqresp(sys, w_range / fs)
f = w / 2 / pi
factor = h[0] / h2[0]
f2 = w2 / 2 / pi * fs
CombinedPlot() \
.setTitle(title) \
.setXTitle("Frecuencia (hz)") \
.setYTitle("Amplitud (Db)") \
.addSignalPlot(
signal=Senial.Senial(
f, 20 * log10(abs(h))
),
color="red",
name="Analógica"
) \
.addSignalPlot(
signal=Senial.Senial(
f2, 20 * log10(abs(h2) * factor)
),
color="blue",
name="Digital método invariante al impulso"
).plot().save("output/" + filename)
