Testing different fibre composites¶
Fabrication methods were the same - a board covered in saran wrap, 3 layers (except the Emmakites 40D Ripstop Nylon which was also tested at 6 layers) of material and epoxy, breather fabric, breather film, all vacuum bagged.
Sample data S15 and S19 were retested as S27 and S28. In addition, S14, S20 and S28 only contain 1 data point, it's thought that the process with which S15 and S19 were halted may have impacted the subsequent/prior tests. Furthermore, the Emmakites data samples seem to be incomplete, judging by the number of data points in each file.
In [82]:
import numpy as np
import csv
from matplotlib import pyplot as plt
cols = ['#cc33ff', '#33cc33', '#6666ff', '#ff0066', '#00cc99', '#0066ff', '#ff0000', '#003399', '#ffcc00', '#990033', '#ff9933', '#cc99ff', '#33cccc']
# sample_no, mass (g), length (mm), width (mm), height (mm), instron data
data = [[1.1, 5.3, 112.1, 35.13, 1.55, 1, 'Medium Weight Linen White'],
[1.2, 4.4, 112.46, 34.64, 0.94,2, 'Medium Weight Linen White'],
[2.1, 5.7, 112.0, 34.44, 1.35, 3, 'Kaufman Brussels Washer Linen Blend Yarn Dye Flax'],
[2.2, 5.6, 111.25, 33.65, 1.22, 4, 'Kaufman Brussels Washer Linen Blend Yarn Dye Flax'],
[3.1, 6.1, 111.31, 34.05, 1.46, 5, 'Kaufman Essex Linen Blend Natural Fabric'],
[3.2, 6.2, 111.89, 33.95, 1.45, 6, 'Kaufman Essex Linen Blend Natural Fabric'],
[4.1, 9.0, 111.68, 34.17, 2.13, 7, '9.3 oz. Canvas Duck Natural'],
[4.2, 8.9, 111.44, 33.72, 1.96, 8, '9.3 oz. Canvas Duck Natural'],
[5.1, 6.2, 111.62, 34.61, 1.37, 9, 'European 100% Linen Stone Fabric'],
[5.2, 7.7, 112.2, 34.39, 2.21, 10, 'European 100% Linen Stone Fabric'],
[6.1, 2.8, 110.98, 33.53, 0.8, 11, 'Emmakites 40D Ripstop Nylon - 3 layers'],
[6.2, 3.0, 112.08, 33.73, 0.95, 12, 'Emmakites 40D Ripstop Nylon - 3 layers'],
[6.3, 2.9, 112.81, 33.02, 0.94, 13, 'Emmakites 40D Ripstop Nylon - 6 layers'],
#[6.4, 2.6, 108.35, 35.00, 0.89, 14, 'Emmakites 40D Ripstop Nylon - 6 layers'],
[7.1, 8.0, 110.92, 33.68, 1.82, 27, 'Organic Cotton Ripstop Natural'],
[7.2, 8.0, 112.63, 34.14, 2.07, 16, 'Organic Cotton Ripstop Natural'],
[8.1, 5.4, 110.41, 33.62, 1.18, 17, 'Natural Linen Fabric, Zakka Cotton/linen, Organic Unbleached'],
[8.2, 5.4, 112.39, 33.97, 1.21, 18, 'Natural Linen Fabric, Zakka Cotton/linen, Organic Unbleached'],
#[9.1, 1.7, 113.86, 33.01, 0.49, 28, 'Black 70 Denier Nylon Ripstop'],
#[9.2, 1.6, 113.35, 32.87, 0.63, 20, 'Black 70 Denier Nylon Ripstop'],
[10.1, 5.9, 112.14, 33.59, 1.36, 21, 'Natural Irish Linen Fabric'],
[10.2, 6.0, 111.62, 34.15, 1.48, 22, 'Natural Irish Linen Fabric'],
[11.1, 5.4, 110.39, 34.04, 1.32, 23, 'European 100% Linen Oatmeal Fabric'],
[11.2, 5.7, 111.80, 35.13, 1.24, 24, 'European 100% Linen Oatmeal Fabric'],
[12.1, 9.6, 112.23, 33.99, 2.14, 25, '10-Ounces Natural Canvas'],
[12.2, 9.2, 112.88, 34.13, 2.21, 26, '10-Ounces Natural Canvas']]
SN = 0; M = 1; L = 2; W = 3; H = 4; ID = 5; MT = 6;
data = np.array(data)
sample_number = np.array(data[:,SN])
length = np.array(data[:,L], dtype=np.float64)
width = np.array(data[:,W], dtype=np.float64)
height = np.array(data[:,H], dtype=np.float64)
mass = np.array(data[:,M], dtype=np.float64)
material_type = np.array(data[:,MT])
volume = np.dot(length, np.dot(width, height)) # in mm^3
density = np.divide(mass,volume)*1000000 # in kg/m^3
float_data = np.column_stack((np.column_stack((np.column_stack((mass,length)),width)),height))
In [83]:
# takes the instron csv files and makes numpy arrays of the data for each sample
# note - S15 and S19 were retested as S27 and S28
# make the string for the csv file of each sample
all_instron_data = []
file_name = "instron_data_2017.05.10/Specimen_RawData_" + str(1) + ".csv"
T = 0; D = 1; F = 2;
# extract the row from the csv file and then pull out the numbers
for h in range(len(data)):
file_name = "instron_data_2017.05.10/Specimen_RawData_" + str(int(data[h][ID])) + ".csv"
with open(file_name, 'rb') as csvfile:
csvreader = csv.reader(csvfile, delimiter=',', quotechar='|')
instron_data_np = []
instron_data_str = []
for row in csvreader:
r = ', '.join(row)
instron_data_str.append(r)
for i in range(2,len(instron_data_str)):
extract = np.fromstring(instron_data_str[i], dtype=float, sep=',')
instron_data_np.append(extract)
all_instron_data.append(instron_data_np)
all_instron_data = np.array(all_instron_data)
print np.shape(all_instron_data)
In [84]:
def prepare_data_to_plot(dims,instron_data):
data_to_plot = []
print(np.shape(instron_data))
for i in range(len(dims)):
test_data = instron_data[i]
specimen_data = dims[i]
l = specimen_data[1]
w = specimen_data[2]
h = specimen_data[3]
test_data = np.array(test_data)
deflection = test_data[:,D]
load = test_data[:,F]
# d/L for the x axis
X = -1*deflection/l
#(F L**2)/(4 w h**3) for y axis
Y = load * (-l**2/(4*w*h**3))
data_to_plot.append([X,Y])
return data_to_plot
def plot_instron_data(d):
t1 = "Instron testing"
font = {'family': 'monospace',
'color': '#39393d',
'weight': 'light',
'size': 14,
}
fig = plt.figure(figsize=(15,15), facecolor="white")
ax = fig.add_subplot(111)
ax.set_title(t1, fontdict=font)
ax.set_xlabel('d/L',fontdict=font)
ax.set_ylabel('FL^2/4wh^3',fontdict=font)
for j in range(len(d)):
#print("Sample " + str(j+1) + " has " + str(len(d[j][0])) + " data points")
print("Sample " + data[j][5] + " has " + str(len(d[j][0])) + " data points")
if j%2==0:
ax.plot(d[j][0],d[j][1],color=cols[j/2],label=material_type[j])
else:
ax.plot(d[j][0],d[j][1],color=cols[j/2])
plt.legend()
plt.show()
plot_data = prepare_data_to_plot(float_data,all_instron_data)
plot_instron_data(plot_data)
# specimen 11, 12 and 13 don't seem to have the full set of extensions?
# sampledata 20 and 28 only contains a single data point - maybe something has happened to the data since I last plotted?
In [85]:
# calculate stiffness
def fit_function(x,y,n):
N = n
M = 2
A = []
x = np.array(x)
A = np.dstack((np.ones(n),x))
U,s,V = np.linalg.svd(A[0], full_matrices=True)
V = V.T
w_inv = np.array([1/wi if wi!=0 else 0 for wi in s])
S = np.zeros((2,n)); S[:2,:2] = np.diag(w_inv)
v = np.dot(V,np.dot(S,U.T))
b = np.dot(v,y)
return b
gradients = []
for y in range(len(plot_data)):
gradients.append(fit_function(plot_data[y][0],plot_data[y][1],len(plot_data[y][1])))
gradients = np.array(gradients)
gradients_with_labels = []
for i in range(len(gradients)):
gradients_with_labels.append([float(gradients[i][1]),data[i][MT]])
gradients_with_labels = np.array(gradients_with_labels)
gradients_sorted = np.sort(gradients_with_labels,axis=0)
gradients_sorted = np.array(gradients_sorted)
# only printing values with a positive gradient - doesn't seem to be sorting properly so
# perhaps the floats aren't being maintained, so the sort method using a list doesn't work,
# I think I need to sort as a dictionary - could have been because some files only had 1 data point
# I've removed them for now
def calc_plot_stiffness():
t1 = "Stiffness values from gradients"
font = {'family': 'monospace',
'color': '#39393d',
'weight': 'light',
'size': 25,
}
fig = plt.figure(figsize=(20,20), facecolor="white")
ax = fig.add_subplot(111)
ax.set_title(t1, fontdict=font)
ax.set_xlabel('Material',fontdict=font)
ax.set_ylabel('Stiffness',fontdict=font)
ind = np.arange(len(gradients_with_labels))
width = 0.25
heights = []
for i in range(len(gradients_with_labels)):
heights.append(gradients_with_labels[i][0])
#labels.append(gradients_with_labels[i][1])
labels = gradients_with_labels[:,1]
rects = ax.bar(ind, heights, width, height, color='#610dab',label=labels)
plt.subplots_adjust(bottom=0.3)
xticks_pos = [0.65*patch.get_width() + patch.get_xy()[0] for patch in rects]
plt.xticks(xticks_pos, labels, ha='right', rotation=55, fontsize=20,fontname='monospace')
plt.show()
# need to check whether the labels actually match the data on the graph
calc_plot_stiffness()
In [ ]: