graph_extruder: Internal updates to graph calculation

Signed-off-by: Kevin O'Connor <kevin@koconnor.net>
This commit is contained in:
Kevin O'Connor 2020-02-18 12:38:45 -05:00
parent 17f2b4e960
commit 0e37f8c9b3
1 changed files with 74 additions and 48 deletions

View File

@ -1,7 +1,7 @@
#!/usr/bin/env python2
# Generate extruder pressure advance motion graphs
#
# Copyright (C) 2019 Kevin O'Connor <kevin@koconnor.net>
# Copyright (C) 2019-2020 Kevin O'Connor <kevin@koconnor.net>
#
# This file may be distributed under the terms of the GNU GPLv3 license.
import math, optparse, datetime
@ -17,7 +17,7 @@ INV_SEG_TIME = 1. / SEG_TIME
# List of moves: [(start_v, end_v, move_t), ...]
Moves = [
(0., 0., .200),
(0., 0., .100),
(0., 100., None), (100., 100., .200), (100., 60., None),
(60., 100., None), (100., 100., .200), (100., 0., None),
(0., 0., .300)
@ -47,85 +47,111 @@ def gen_positions():
start_t = end_t
return out
######################################################################
# List helper functions
######################################################################
MARGIN_TIME = 0.050
def time_to_index(t):
return int(t * INV_SEG_TIME + .5)
def indexes(positions):
drop = time_to_index(MARGIN_TIME)
return range(drop, len(positions)-drop)
def trim_lists(*lists):
keep = len(lists[0]) - time_to_index(2. * MARGIN_TIME)
for l in lists:
del l[keep:]
######################################################################
# Common data filters
######################################################################
# Generate estimated first order derivative
def gen_deriv(data):
return [0.] + [(data[i+1] - data[i]) * INV_SEG_TIME
for i in range(len(data)-1)]
def time_to_index(t):
return int(t * INV_SEG_TIME + .5)
# Simple average between two points smooth_time away
def calc_average(positions, smooth_time):
offset = time_to_index(smooth_time * .5)
out = [0.] * len(positions)
for i in indexes(positions):
out[i] = .5 * (positions[i-offset] + positions[i+offset])
return out
# Average (via integration) of smooth_time range
def calc_smooth(positions, smooth_time):
offset = time_to_index(smooth_time * .5)
weight = 1. / (2*offset - 1)
out = [0.] * len(positions)
for i in indexes(positions):
out[i] = sum(positions[i-offset+1:i+offset]) * weight
return out
# Time weighted average (via integration) of smooth_time range
def calc_weighted(positions, smooth_time):
offset = time_to_index(smooth_time * .5)
weight = 1. / offset**2
out = [0.] * len(positions)
for i in indexes(positions):
weighted_data = [positions[j] * (offset - abs(j-i))
for j in range(i-offset, i+offset)]
out[i] = sum(weighted_data) * weight
return out
######################################################################
# Pressure advance
######################################################################
PA_HALF_SMOOTH_T = .040 / 2.
SMOOTH_TIME = .040
PRESSURE_ADVANCE = .045
# Calculate raw pressure advance positions
def calc_pa_raw(t, positions):
def calc_pa_raw(positions):
pa = PRESSURE_ADVANCE * INV_SEG_TIME
i = time_to_index(t)
return positions[i] + pa * (positions[i+1] - positions[i])
out = [0.] * len(positions)
for i in indexes(positions):
out[i] = positions[i] + pa * (positions[i+1] - positions[i])
return out
# Pressure advance smoothed using average velocity (for reference only)
def calc_pa_average(t, positions):
pa_factor = PRESSURE_ADVANCE / (2. * PA_HALF_SMOOTH_T)
base_pos = positions[time_to_index(t)]
start_pos = positions[time_to_index(t - PA_HALF_SMOOTH_T)]
end_pos = positions[time_to_index(t + PA_HALF_SMOOTH_T)]
return base_pos + (end_pos - start_pos) * pa_factor
# Pressure advance with simple time smoothing (for reference only)
def calc_pa_smooth(t, positions):
start_index = time_to_index(t - PA_HALF_SMOOTH_T) + 1
end_index = time_to_index(t + PA_HALF_SMOOTH_T)
pa = PRESSURE_ADVANCE * INV_SEG_TIME
pa_data = [positions[i] + pa * (positions[i+1] - positions[i])
for i in range(start_index, end_index)]
return sum(pa_data) / (end_index - start_index)
# Calculate pressure advance smoothed using a "weighted average"
def calc_pa_weighted(t, positions):
base_index = time_to_index(t)
start_index = time_to_index(t - PA_HALF_SMOOTH_T) + 1
end_index = time_to_index(t + PA_HALF_SMOOTH_T)
diff = .5 * (end_index - start_index)
pa = PRESSURE_ADVANCE * INV_SEG_TIME
pa_data = [(positions[i] + pa * (positions[i+1] - positions[i]))
* (diff - abs(i-base_index))
for i in range(start_index, end_index)]
return sum(pa_data) / diff**2
# Pressure advance after smoothing
def calc_pa(positions):
return calc_weighted(calc_pa_raw(positions), SMOOTH_TIME)
######################################################################
# Plotting and startup
######################################################################
MARGIN_TIME = 0.100
def plot_motion():
# Nominal motion
positions = gen_positions()
drop = int(MARGIN_TIME * INV_SEG_TIME)
times = [SEG_TIME * t for t in range(len(positions))][drop:-drop]
velocities = gen_deriv(positions[drop:-drop])
velocities = gen_deriv(positions)
accels = gen_deriv(velocities)
# Motion with pressure advance
pa_positions = [calc_pa_raw(t, positions) for t in times]
pa_positions = calc_pa_raw(positions)
pa_velocities = gen_deriv(pa_positions)
# Smoothed motion
sm_positions = [calc_pa_weighted(t, positions) for t in times]
sm_positions = calc_pa(positions)
sm_velocities = gen_deriv(sm_positions)
# Build plot
shift_times = [t - MARGIN_TIME for t in times]
times = [SEG_TIME * i for i in range(len(positions))]
trim_lists(times, velocities, accels,
pa_positions, pa_velocities,
sm_positions, sm_velocities)
fig, ax1 = matplotlib.pyplot.subplots(nrows=1, sharex=True)
ax1.set_title("Extruder Velocity")
ax1.set_ylabel('Velocity (mm/s)')
pa_plot, = ax1.plot(shift_times, pa_velocities, 'r',
pa_plot, = ax1.plot(times, pa_velocities, 'r',
label='Pressure Advance', alpha=0.3)
nom_plot, = ax1.plot(shift_times, velocities, 'black', label='Nominal')
sm_plot, = ax1.plot(shift_times, sm_velocities, 'g', label='Smooth PA',
alpha=0.9)
nom_plot, = ax1.plot(times, velocities, 'black', label='Nominal')
sm_plot, = ax1.plot(times, sm_velocities, 'g', label='Smooth PA', alpha=0.9)
fontP = matplotlib.font_manager.FontProperties()
fontP.set_size('x-small')
ax1.legend(handles=[nom_plot, pa_plot, sm_plot], loc='best', prop=fontP)