-
Notifications
You must be signed in to change notification settings - Fork 1
/
Battery7_RULforStage3_B5.m
170 lines (154 loc) · 5.79 KB
/
Battery7_RULforStage3_B5.m
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
clc; clear; close all
%% Load data-Extract the discharging cycles from Battery 5 and 7
load('data/Data_bettery.mat')
DATA_B7 = DATA_No7; Capacity_B7 = Capacity_No7;
DATA_B5 = DATA_No5; Capacity_B5 = Capacity_No5;
DATA_B5(43) = []; Capacity_B5(43) = [];
Phase1 = [1, 30]; Phase2 = [31, 107]; Phase3 = [108, 165];
%% Calculate the minimum length of each phase
[TrainLength3, Trainsample3] = CalLength(DATA_B7, Phase3, 0.7);
[TrainLength1, Trainsample1] = CalLength(DATA_B5, Phase3, 0.7);
TrainLength = min([TrainLength1, TrainLength3]);
Trainsample = min([Trainsample1, Trainsample3]);
%% Obtain the staionary and non-staionary source from training data
tau = 5; m = 3; s = 4;
[TrData_B7, XtrainB7, YS] = Dataslicing1(DATA_B7, Capacity_B7, Phase3, TrainLength, Trainsample, tau, m);
Xtrain = XtrainB7;
[est_Ps, est_Pn, est_As, est_An, ssa_results] = ssa(Xtrain, s, 'reps', 20, 'equal_epochs', Trainsample, 'random_seed', 12345);
Ts = est_Ps * Xtrain; Tn = est_Pn * Xtrain;
[Tn,meanX,stdX] = autoscale_new(Tn');
%% Arranging the training dataset
Xtrain = []; Xtrain_error = [];
for d = Phase3(1):Phase3(1)+Trainsample-1
xtrain = [];
N = size(TrData_B7{d},1);
xtrain = mapminmax('apply', TrData_B7{d}', YS);
xtrain_error = est_Ps * xtrain;
xtrain = est_Pn * xtrain;
for j = 1:9-s
xtrain(j,:) = smoothdata(xtrain(j,:),'gaussian',15);
end
for j = 1:s
xtrain_error(j,:) = smoothdata(xtrain_error(j,:),'gaussian',11);
end
xtrain = autoscale_new(xtrain',meanX,stdX);
xtrain = xtrain';
xtrain(:,231:end) = [];
xtrain_error(:,231:end) = [];
xtrain_error(:,1:20) = [];
N = size(xtrain,2);
cycleID = 1:N;
Capacity = Capacity_B7(d)*ones(N,1);
TrData_B7{d} = [(d-Phase3(1)+1)*ones(N,1) cycleID' xtrain' Capacity];
Xtrain = [Xtrain; TrData_B7{d}];
Xtrain_error{d-Phase3(1)+1} = xtrain_error;
end
%% Arranging Testing Dataset
Xtest = []; TT = []; Xtest_D = []; Xtest_Error = [];
for d = Phase3(1):Phase3(2)
xtest = [];
N = size(DATA_B5{d},2);
data = DATA_B5{d}(:,1:TrainLength);
N = size(data, 2);
data = [reconstitution(data(1,:), N, m, tau); reconstitution(data(2,:), N, m, tau); reconstitution(data(3,:), N, m, tau)];
data = mapminmax('apply', data, YS);
xtest_error = est_Ps * data;
xtest = est_Pn * data;
for j = 1:9-s
xtest(j,:) = smoothdata(xtest(j,:),'gaussian',15);
end
for j = 1:s
xtest_error(j,:) = smoothdata(xtest_error(j,:),'gaussian',11);
end
xtest(:,231:end) = [];
xtest_error(:,231:end) = [];
xtest_error(:,1:20) = [];
xtest = autoscale_new(xtest',meanX,stdX);
xtest = xtest';
cycleID = 1:size(xtest,2);
Xtest{d-Phase3(1)+1} = [(d-Phase3(1)+1)*ones(size(xtest,2),1) cycleID' xtest'];
Xtest_D = [Xtest_D; Xtest{d-Phase3(1)+1}];
TT = [TT xtest];
Xtest_Error{d-Phase3(1)+1} = xtest_error;
end
%% Arrange the data for the following LSTM
for d = Phase3(1):Phase3(1)+Trainsample-1
TrData_B7{d}(:, 1:2) = [];
TrData_B7{d}(:, end) = [];
TrData_B7{d} = TrData_B7{d}';
end
TrData = TrData_B7(Phase3(1):Phase3(1)+Trainsample-1);
for d = 1: Phase3(2)-Phase3(1)+1
Xtest{d}(:, 1:2) = [];
Xtest{d} = Xtest{d}';
end
X_training = TrData; Y_training = Capacity_No7(Phase3(1):Phase3(1)+Trainsample-1);
X_validate = Xtest; Y_validate = Capacity_No5(Phase3(1):Phase3(2));
%% Train the model
inputsize = 9-s;
numHiddenUnits1 = 100; numHiddenUnits2 = 100; numResponses = 1;
layers = [ ...
sequenceInputLayer(inputsize)
lstmLayer(numHiddenUnits1, 'OutputMode', 'sequence')
lstmLayer(numHiddenUnits2, 'OutputMode', 'last')
fullyConnectedLayer(50)
fullyConnectedLayer(numResponses)
regressionLayer];
maxEpochs = 200; miniBatchSize = 4;
options = trainingOptions('adam', ...
'MaxEpochs', maxEpochs, ...
'MiniBatchSize', miniBatchSize, ...
'InitialLearnRate', 0.01, ...
'GradientThreshold', 1, ...
'Shuffle', 'every-epoch', ...
'Plots', 'training-progress',...
'Verbose', 0);
net = trainNetwork(X_training, Y_training', layers, options);
%% Plot the training data
YPred_train =[]; Ytrain = [];
for i = 1 : size(X_training, 2)
yPred_train = predict(net, X_training{i}, 'MiniBatchSize',1);
YPred_train = [YPred_train yPred_train];
Ytrain = [Ytrain Y_training(i)];
end
YPred_validate = []; Yvalidate = [];
for i = 1 : size(X_validate, 2)
yPred_validate = predict(net, X_validate{i}, 'MiniBatchSize',1);
YPred_validate = [YPred_validate yPred_validate];
Yvalidate = [Yvalidate Y_validate(i)];
end
%% Plot the concerned information for training dataset
figure
subplot(121)
plot(YPred_train, 'o-', 'Linewidth', 1.5, 'MarkerSize', 4);
hold on;
plot(Ytrain, 'h--', 'MarkerFaceColor', 'r');
xlabel("No. of training cycles");
ylabel("Capacity (Amp-hr)");
legend('Prediction (Proposed)', 'Real data');
xlim([0, size(Ytrain, 2)]);
subplot(122)
bar(YPred_train-Ytrain)
xlabel("No. of training cycles")
ylabel("Error (Amp-hr)")
xlim([0, size(Ytrain, 2)])
%% Plot the concerned information for testing dataset
K = 1:size(Yvalidate, 2);
figure
subplot(121)
plot(YPred_validate, 'o-','Linewidth', 1.5, 'MarkerSize',4)
hold on
plot(Yvalidate, 'h--', 'MarkerFaceColor', 'r')
xlabel("No. of testing cycles")
ylabel("Capacity (Amp-hr)");
legend('Prediction (Proposed)','Real data');
xlim([0, size(Yvalidate, 2)])
subplot(122)
bar(YPred_validate-Yvalidate)
xlabel("No. of testing cycles")
ylabel("Error (Amp-hr)")
xlim([0, size(Yvalidate, 2)])
%% Calculate the estimation error
m = mean(YPred_validate-Yvalidate);
s = std(YPred_validate-Yvalidate);
RMSE = sqrt(mean((YPred_validate-Yvalidate).^2))/2*100