-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathlevel2_plot.py
148 lines (131 loc) · 7.33 KB
/
level2_plot.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
import pandas as pd
import matplotlib.pyplot as plt
import numpy as np
#--------------- Style ---------------------
plt.rcParams['font.monospace'] = 'Ubuntu Mono'
plt.rcParams['font.size'] = 15
plt.rcParams['axes.labelsize'] = 14
plt.rcParams['axes.labelweight'] = 'bold'
plt.rcParams['xtick.labelsize'] = 14
plt.rcParams['ytick.labelsize'] = 14
plt.rcParams['legend.fontsize'] = 15
plt.rcParams['legend.fancybox'] = True
plt.rcParams['legend.shadow'] = True
plt.rcParams['lines.linewidth'] = 2
# -------------------------------Scenario comparison---------------------------------
# Import data
def scenario():
in_out = pd.read_excel("Data/level-2-4-case-comparison-730K.xlsx", usecols='B:C', skiprows=3, nrows=98).to_numpy()
rec_only = pd.read_excel("Data/level-2-4-case-comparison-730K.xlsx", usecols='E:F', skiprows=2, nrows=98).to_numpy()
pmfuel = pd.read_excel("Data/level-2-4-case-comparison-730K.xlsx", usecols='H:I', skiprows=2, nrows=98).to_numpy()
mefuel = pd.read_excel("Data/level-2-4-case-comparison-730K.xlsx", usecols='K:L', skiprows=2, nrows=98).to_numpy()
ep1 = 32.5528301886792
# Plot Graph
fig, ax = plt.subplots(2, 1)
ax[1].plot(in_out[:,1], in_out[:,0]/1000000, label="In/Out (Case 1)", color="darkorange", linewidth=3.5)
ax[0].plot(rec_only[:,1], rec_only[:,0]/1000000, label="Recycle Only (Case 2)", color="forestgreen", linewidth=3.5)
ax[1].plot(rec_only[:,1], rec_only[:,0]/1000000, label="Recycle Only (Case 2)", color="forestgreen", linewidth=3.5)
ax[1].plot(mefuel[:,1], mefuel[:,0]/1000000, label = "Toluene Recycle + Methanol Fuel (Case 3)", color = "crimson", linewidth=3.5)
ax[0].plot(pmfuel[:,1], pmfuel[:,0]/1000000, label="Recycle + Xylene Fuel (Case 4)", color = "royalblue", linewidth=3.5)
ax[1].plot(pmfuel[:, 1], pmfuel[:,0]/1000000, label="Recycle + Xylene Fuel (Case 4)", color="royalblue", linewidth=3.5)
ax[0].set_ylim([0, 40])
ax[0].set_xlim([0.025, 0.35])
ax[0].set_xticklabels([])
ax[0].xaxis.set_tick_params(labelsize=25)
ax[0].yaxis.set_tick_params(labelsize=25)
ax[1].xaxis.set_tick_params(labelsize=25)
ax[1].yaxis.set_tick_params(labelsize=25)
ax[1].set_ylim([-1000, 0])
ax[1].set_xlim([0.025, 0.35])
ax[0].axhline(ep1, linestyle='--', color='black', label='EP1', linewidth=3.5)
ax[1].axhline(ep1, linestyle='--', color='black', label='EP1')
ax[1].legend(loc='upper center', bbox_to_anchor=(0.5, 2.3), ncol=3, fancybox=True, shadow=True, fontsize=25)
fig.text(0.04, 0.5, "Economic Potential (Million USD/yr)", va='center', rotation='vertical', fontsize=30, fontweight='bold')
fig.text(0.5, 0.04, "Conversion", ha='center', fontsize=30, fontweight='bold')
plt.subplots_adjust(hspace=0)
plt.show()
# -------------------------------Temperature comparison---------------------------------
# Import data
def temperature():
temp_730 = pd.read_excel("Data/level-2-temperature-comparison-final.xlsx", sheet_name="Sheet5", usecols='B:C', skiprows=2).to_numpy()
temp_760 = pd.read_excel("Data/level-2-temperature-comparison-final.xlsx", sheet_name="Sheet5", usecols='E:F', skiprows=2).to_numpy()
temp_790 = pd.read_excel("Data/level-2-temperature-comparison-final.xlsx", sheet_name="Sheet5", usecols='H:I', skiprows=2).to_numpy()
temp_820 = pd.read_excel("Data/level-2-temperature-comparison-final.xlsx", sheet_name="Sheet5", usecols='K:L', skiprows=2).to_numpy()
#Plot Graph
plt.plot(temp_730[:,1], temp_730[:,0]/1000000, label="730K")
plt.plot(temp_760[:,1], temp_760[:,0]/1000000, label="760K")
plt.plot(temp_790[:,1], temp_790[:,0]/1000000, label="790K")
plt.plot(temp_820[:,1], temp_820[:,0]/1000000, label="820K")
plt.hlines(0,0,100,linestyles="--")
#plt.title("EP Comparison of Different Temperatures")
plt.ylabel("Economic Potential (Million USD/yr)")
plt.xlabel("Conversion")
plt.legend()
plt.ylim([-40, 40])
plt.xlim([0, 0.43])
plt.show()
# -------------------------------Pressure comparison---------------------------------
# Import data
def pressure():
pres_1 = pd.read_excel("Data/level 2 plots.xlsx", sheet_name="Pressure comparison", usecols='B:C', skiprows=2, nrows=97).to_numpy()
pres_5 = pd.read_excel("Data/level 2 plots.xlsx", sheet_name="Pressure comparison", usecols='E:F', skiprows=2, nrows=99).to_numpy()
pres_10 = pd.read_excel("Data/level 2 plots.xlsx", sheet_name="Pressure comparison", usecols='H:I', skiprows=2, nrows=99).to_numpy()
#Plot Graph
plt.plot(pres_1[:,1], pres_1[:,0]/1000000, label="1 bar")
plt.plot(pres_5[:,1], pres_5[:,0]/1000000, label="5 bar")
plt.plot(pres_10[:,1], pres_10[:,0]/1000000, label="10 bar")
plt.hlines(0,0,100,linestyles="--")
#plt.title("EP Comparison of Different Pressures")
plt.ylabel("Economic Potential (Million USD/yr)")
plt.xlabel("Conversion")
plt.legend()
plt.xlim([0.05, 0.45])
plt.ylim([-50, 40])
plt.show()
'''
# -------------------------------Feed Ratio comparison---------------------------------
# Import data
fr1 = pd.read_excel("Data/level 2 plots.xlsx", sheet_name="Pressure comparison", usecols='B:C', skiprows=2, nrows=97).to_numpy()
fr2 = pd.read_excel("Data/level 2 plots.xlsx", sheet_name="Pressure comparison", usecols='E:F', skiprows=2, nrows=99).to_numpy()
fr3 = pd.read_excel("Data/level 2 plots.xlsx", sheet_name="Pressure comparison", usecols='H:I', skiprows=2, nrows=99).to_numpy()
#Plot Graph
plt.figure(3)
plt.plot(pres_1[:,1], pres_1[:,0]/1000000, label="1 bar")
plt.plot(pres_5[:,1], pres_5[:,0]/1000000, label="5 bar")
plt.plot(pres_10[:,1], pres_10[:,0]/1000000, label="10 bar")
plt.title("EP Comparison of Different Pressures")
plt.ylabel("Economic Potential (Million USD/yr)")
plt.xlabel("Conversion")
plt.legend()
plt.xlim([0.05, 0.45])
plt.ylim([-50, 40])
plt.show()
'''
# -------------------------------Feed Ratio comparison with EP---------------------------------
def feed_ratio_ep():
one = pd.read_excel("Data/LEVEL 3 excel.xlsx", sheet_name="EP2 vary feed ratio", usecols='A:B', nrows=20).to_numpy()
two = pd.read_excel("Data/LEVEL 3 excel.xlsx", sheet_name="EP2 vary feed ratio", usecols='A:B', nrows=13, skiprows=23).to_numpy()
three = pd.read_excel("Data/LEVEL 3 excel.xlsx", sheet_name="EP2 vary feed ratio", usecols='A:B', nrows=11, skiprows=39).to_numpy()
four = pd.read_excel("Data/LEVEL 3 excel.xlsx", sheet_name="EP2 vary feed ratio", usecols='A:B', nrows=10, skiprows=53).to_numpy()
plt.plot(one[:, 0] / 100, one[:, 1] / 1000000, label="MeOH 1:1 Tol")
plt.plot(two[:, 0] / 100, two[:, 1] / 1000000, label="MeOH 1:2 Tol")
plt.plot(three[:, 0] / 100, three[:, 1] / 1000000, label="MeOH 1:3 Tol")
plt.plot(four[:, 0] / 100, four[:, 1] / 1000000, label="MeOH 1:4 Tol")
plt.hlines(0,0,100,linestyles="--")
plt.ylabel("Economic Potential (Million USD/yr)")
plt.xlabel("Conversion")
plt.xlim([0.025, 0.25])
plt.ylim([-45, 35])
plt.legend()
plt.show()
#------------------------------Selectivity vs Conversion------------------------------
def sel_conv():
df = pd.read_excel("Data/level 2 selec vs convo.xlsx", sheet_name="Sheet2", usecols='B:I', skiprows=1).to_numpy()
plt.plot(df[:, 0], df[:, 1], label="730K")
plt.plot(df[:, 6], df[:, 7], label="750K")
plt.plot(df[:, 3], df[:, 4], label="770K")
plt.ylabel("Selectivity")
plt.xlabel("Conversion")
plt.legend(loc='lower left')
plt.show()
scenario()