From 349f33b45612a17648de7166d99d7766a7c382eb Mon Sep 17 00:00:00 2001
From: Junjie <junjie.zhang@eonerc.rwth-aachen.de>
Date: Wed, 7 Aug 2019 20:27:52 +0200
Subject: [PATCH] add instFrequency() method to timeSeries
---
villas/dataprocessing/timeseries.py | 26 +++++++++++++++++++++++---
1 file changed, 23 insertions(+), 3 deletions(-)
diff --git a/villas/dataprocessing/timeseries.py b/villas/dataprocessing/timeseries.py
index ff2276c..91583d1 100644
--- a/villas/dataprocessing/timeseries.py
+++ b/villas/dataprocessing/timeseries.py
@@ -1,5 +1,6 @@
import numpy as np
import cmath
+from scipy.signal import hilbert, chirp
class TimeSeries:
"""Stores data from different simulation sources.
@@ -44,6 +45,15 @@ class TimeSeries:
ts_real = TimeSeries(self.name+'_real', self.time, _real)
return ts_real
+ def imag(self):
+ """ get the imaginary part of complex time series.
+ """
+ _imag = []
+ for value in self.values:
+ _imag.append(np.imag(value))
+ ts_imag = TimeSeries(self.name+'_real', self.time, _imag)
+ return ts_imag
+
def phasor(self):
"""Calculate phasors of complex time series
and return dict with absolute value and phase.
@@ -199,8 +209,8 @@ class TimeSeries:
"""Returns relative difference between two time series objects to the first.
"""
diff_val=TimeSeries.diff('diff', ts1, ts2).values
- # in case an element in ts1 is zero and the corresponding diff is non-zero
- rel_diff_to_ts1=np.divide(diff_val, ts1.values, out=np.zeros_like(diff_val), where=ts1.values!=0)
+ # relative error to the max value of ts1
+ rel_diff_to_ts1 = diff_val/ts1.values.max()
ts_rel_diff_to_ts1 = TimeSeries(name, ts1.time, rel_diff_to_ts1)
return ts_rel_diff_to_ts1
@@ -222,5 +232,15 @@ class TimeSeries:
phasor_list = {}
for name, ts in timeseries_list.items():
phasor_list[name] = ts.phasor()
+ return phasor_list
- return phasor_list
\ No newline at end of file
+ @staticmethod
+ def instFrequency(ts):
+ duration = ts.time.max()
+ samples = ts.time.size
+ sampling_freq = samples / duration
+ analytic_signal = hilbert(ts.abs().values)
+ instaneous_phase=np.unwrap(np.angle(analytic_signal))
+ instantaneous_frequency = (np.diff(instaneous_phase) / (2.0 * np.pi) * sampling_freq)
+ ts_instFreq=TimeSeries(str(ts.name + "instFreq"), ts.time[:-1], instantaneous_frequency)
+ return ts_instFreq
--
GitLab