-
Notifications
You must be signed in to change notification settings - Fork 2
/
Copy pathenv.py
203 lines (167 loc) · 7.53 KB
/
env.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
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
import copy
from batterymodel.battery import Battery
from gym import Env
import numpy as np
from gym import spaces
class Env(Env):
def __init__(self, time_passed_by_step = 1, randomization = True, scenario = 0, charging = True, num_cells = 4, length_reference_run = 2600):
self.randomization = randomization
self.scenario = scenario
self.discharging_current_for_evaluation = [3.95, 4.11, 3.49, 1.82, 2.72, 3.21, 4.73, 5.8, 3.12, 1.8]
self.discharging_current_for_evaluation_counter = 1
self.num_cells = num_cells
self.time_passed_by_step = time_passed_by_step # unit: seconds
self.length_reference_run = length_reference_run
self.charging = charging
if self.charging:
self.I_BCS = -3
elif self.randomization:
self.I_BCS = np.random.rand() * 6
else:
self.I_BCS = self.discharging_current_for_evaluation[0]
self.m_env = Battery(self.num_cells, self.time_passed_by_step, self.randomization, self.scenario, self.charging)
self.done = False
self.observation_space = spaces.Box(low=0, high=1,shape=([self.num_cells * 2]), dtype=np.float32)
#first action: which cell to choose (needs to be discretized)
#second action: BMS current (needs to be scaled by 2)
self.action_space = spaces.Box(low=-1,high=1,shape=([2]),dtype=np.float32)
self.counter = 0
self.init_diffs()
self.list_SoC_reward = []
self.list_temp_reward =[]
self.list_speed_reward = []
def init_diffs(self):
SoCs = []
Temps = []
for c in self.m_env.get_cells():
SoCs.append(c.SoC)
Temps.append(c.T_cel / 100)
self.best_diff_SoC = self.score_for_closeness_of_values(SoCs,self.num_cells)
self.best_diff_Temp = self.score_for_closeness_of_values(Temps,self.num_cells)
def reset(self):
self.m_env = Battery(self.num_cells, self.time_passed_by_step, self.randomization, self.scenario, self.charging)
self.counter = 0
self.done = False
observation = np.array(self.m_env.get_cell_state(), dtype=np.float32)
self.list_SoC_reward = []
self.list_temp_reward =[]
self.list_speed_reward = []
self.init_diffs()
if self.charging:
self.I_BCS = -3
elif self.randomization:
self.I_BCS = np.random.rand() * 6
else:
self.I_BCS = self.discharging_current_for_evaluation[0]
self.discharging_current_for_evaluation_counter = 1
return observation
def calc_reward(self):
SoCs = []
Temps = []
for c in self.m_env.get_cells():
SoCs.append(c.SoC)
Temps.append(c.T_cel / 100)
diff_SoC = self.score_for_closeness_of_values(SoCs,self.num_cells)
diff_Temp = self.score_for_closeness_of_values(Temps,self.num_cells)
SoC_balancing_reward = self.calculate_reward(diff_SoC,self.best_diff_SoC,0.0075,0.65)
temp_balancing_reward = self.calculate_reward(diff_Temp,self.best_diff_Temp,0.0001,0.65)
speed_reward = -1
return SoC_balancing_reward, temp_balancing_reward, speed_reward
def step(self, action):
self.counter += 1
self.I_BCS = self.calc_I_BCS()
cell_from_action = self.calc_cell(action)
BMS_current = self.calc_I_BMS(action)
done = self.m_env.step(cell_from_action,BMS_current,self.I_BCS)
observation = np.array(self.m_env.get_cell_state(), dtype=np.float32)
SoC_balancing_reward, temp_balancing_reward, speed_reward = self.calc_reward()
if self.counter < self.length_reference_run:
factor_speed_reward = 0.0
elif self.counter < 2 * self.length_reference_run:
factor_speed_reward = (self.counter - self.length_reference_run) * (0.4 / self.length_reference_run)
else:
factor_speed_reward = 0.4
reward = (0.8 * SoC_balancing_reward + 0.2 * temp_balancing_reward) * (1 - factor_speed_reward) + factor_speed_reward * speed_reward
if BMS_current == 0:
reward -= 0.3
elif BMS_current < 0.0001 and BMS_current > -0.0001:
reward -= 0.1
elif BMS_current < 0.1 and BMS_current > -0.1:
reward -= 0.1
self.list_SoC_reward.append(SoC_balancing_reward)
self.list_temp_reward.append(temp_balancing_reward)
self.list_speed_reward.append(speed_reward)
if reward < -1:
reward = -1
if np.array(self.m_env.get_SoC_values()).max() > 0.95 or np.array(self.m_env.get_SoC_values()).min() < 0.1:
self.done = True
return observation, reward, self.done, {}
def calc_I_BCS(self):
if not self.charging and self.counter % (300 / self.time_passed_by_step) == 0:
if self.randomization:
I_BCS = np.random.rand() * 6
else:
I_BCS = self.discharging_current_for_evaluation[self.discharging_current_for_evaluation_counter]
self.discharging_current_for_evaluation_counter = (self.discharging_current_for_evaluation_counter + 1) % 10
if self.charging:
factor_I = 1
if np.array(self.m_env.get_SoC_values()).mean() > 0.3 and np.array(self.m_env.get_SoC_values()).mean() <= 0.45:
factor_I = 0.85
elif np.array(self.m_env.get_SoC_values()).mean() > 0.45 and np.array(self.m_env.get_SoC_values()).mean() <= 0.6:
factor_I = 0.75
elif np.array(self.m_env.get_SoC_values()).mean() > 0.6 and np.array(self.m_env.get_SoC_values()).mean() <= 0.75:
factor_I = 0.6
elif np.array(self.m_env.get_SoC_values()).mean() > 0.75:
factor_I = 0.4
I_BCS = -3 * factor_I
return I_BCS
def calc_cell(self,action):
cell_float = action[0]
if cell_float >= 1.0:
return int(self.num_cells - 1)
elif cell_float < -1.0:
return int(0)
else:
return int(((cell_float + 1) / 2) * self.num_cells)
def calc_I_BMS(self, action):
if self.charging:
BMS_current = ((action[1] + 1) * 3) - 2
else:
BMS_current = (((action[1] + 1) / 2) * 7) - 4
if BMS_current + self.I_BCS > 6:
BMS_current = 6 - self.I_BCS
elif BMS_current + self.I_BCS < -1:
BMS_current = - 1 - self.I_BCS
return BMS_current
def calculate_reward(self,diff,th,plv,nlv):
if th < plv:
th = plv
if diff < plv:
return 1 - (0.1 * (diff/plv))
elif diff == plv:
return 0.9
elif diff < th:
return ((th - diff) / (th - plv)) * 0.9
elif diff == th:
return 0
elif diff < 0.75 - nlv:
return (diff - th) * (- 0.8) * (1 / (0.75 - nlv - th))
elif diff == 0.75 - nlv:
return - 0.8
elif diff < 0.75:
return -0.8 - (diff - (0.75 - nlv)) * (0.2 / (nlv))
else:
return -1
def score_for_closeness_of_values(self,values, num_values):
mean = 0
for v in values:
mean += v
mean = mean / num_values
mean_diff_from_mean = 0
for v in values:
mean_diff_from_mean += abs(mean - v)
mean_diff_from_mean /= num_values
min = np.array(values).min()
max = np.array(values).max()
abs_diff_min_max = max - min
return 0.5 * mean_diff_from_mean + 0.5 * abs_diff_min_max