热传导反问题 matlab,Matlab关于网络反归一化问题

n=[32.6 0.8584 9.37 18 0.14 0 52.1 0.06 7.51 0.6 10.76 3.79 11.74 0.27 1.07 12.54 55 1 0.0377;

32.6 0.8584 9.37 18 0.14 0 52.1 0.06 7.51 0.6 10.76 3.79 11.74 0.27 1.07 12.54 60 1 0.0389;

32.6 0.8584 9.37 18 0.14 0 52.1 0.06 7.51 0.6 10.76 3.79 11.74 0.27 1.07 12.54 65 1 0.0413;

32.6 0.8584 9.37 18 0.14 0 52.1 0.06 7.51 0.6 10.76 3.79 11.74 0.27 1.07 12.54 70 1 0.0772;

32.6 0.8584 9.37 18 0.14 0 52.1 0.06 7.51 0.6 10.76 3.79 11.74 0.27 1.07 12.54 75 1 0.0974;

20.7 0.8894 38.56 -10 0.31 0.025 34.35 0.54 8.03 0 8.8 4.59 21.7 0.08 2.59 29.2 50 1 0.0373;

20.7 0.8894 38.56 -10 0.31 0.025 34.35 0.54 8.03 0 8.8 4.59 21.7 0.08 2.59 29.2 60 1 0.0391;

20.7 0.8894 38.56 -10 0.31 0.025 34.35 0.54 8.03 0 8.8 4.59 21.7 0.08 2.59 29.2 65 1 0.0405;

20.7 0.8894 38.56 -10 0.31 0.025 34.35 0.54 8.03 0 8.8 4.59 21.7 0.08 2.59 29.2 70 1 0.0721;

20.7 0.8894 38.56 -10 0.31 0.025 34.35 0.54 8.03 0 8.8 4.59 21.7 0.08 2.59 29.2 75 1 0.0897;

20.7 0.8894 38.56 -10 0.31 0.025 34.35 0.54 8.03 0 8.8 4.59 21.7 0.08 2.59 29.2 80 1 0.0932;

20.4 0.9276 43.06 -20 0.66 0.12 47.7 0.75 11.21 1.07 2.42 5.9 33.58 0.75 21.17 9.5 50 1 0.0381;

20.4 0.9276 43.06 -20 0.66 0.12 47.7 0.75 11.21 1.07 2.42 5.9 33.58 0.75 21.17 9.5 55 1 0.0467;

20.4 0.9276 43.06 -20 0.66 0.12 47.7 0.75 11.21 1.07 2.42 5.9 33.58 0.75 21.17 9.5 65 1 0.0782;

20.4 0.9276 43.06 -20 0.66 0.12 47.7 0.75 11.21 1.07 2.42 5.9 33.58 0.75 21.17 9.5 70 1 0.093;

20.4 0.9276 43.06 -20 0.66 0.12 47.7 0.75 11.21 1.07 2.42 5.9 33.58 0.75 21.17 9.5 75 1 0.1278;

20.4 0.9276 43.06 -20 0.66 0.12 47.7 0.75 11.21 1.07 2.42 5.9 33.58 0.75 21.17 9.5 80 1 0.1496;

28.9 0.8782 13.65 13 1.12 0 11.5 0.38 0.16 1.06 8.67 3.12 70.73 0.07 46.64 1.53 50 1 0.0439;

28.9 0.8782 13.65 13 1.12 0 11.5 0.38 0.16 1.06 8.67 3.12 70.73 0.07 46.64 1.53 55 1 0.0512;

28.9 0.8782 13.65 13 1.12 0 11.5 0.38 0.16 1.06 8.67 3.12 70.73 0.07 46.64 1.53 60 1 0.0759;

28.9 0.8782 13.65 13 1.12 0 11.5 0.38 0.16 1.06 8.67 3.12 70.73 0.07 46.64 1.53 70 1 0.0992;

28.9 0.8782 13.65 13 1.12 0 11.5 0.38 0.16 1.06 8.67 3.12 70.73 0.07 46.64 1.53 75 1 0.1347;

28.9 0.8782 13.65 13 1.12 0 11.5 0.38 0.16 1.06 8.67 3.12 70.73 0.07 46.64 1.53 80 1 0.162;

33.9 0.8518 7.987 -29 1.15 0.025 90 0.47 0.57 0.28 4.25 5.29 5.87 0.02 6.86 1.05 50 1 0.0448;

33.9 0.8518 7.987 -29 1.15 0.025 90 0.47 0.57 0.28 4.25 5.29 5.87 0.02 6.86 1.05 55 1 0.0688;

33.9 0.8518 7.987 -29 1.15 0.025 90 0.47 0.57 0.28 4.25 5.29 5.87 0.02 6.86 1.05 60 1 0.0864;

33.9 0.8518 7.987 -29 1.15 0.025 90 0.47 0.57 0.28 4.25 5.29 5.87 0.02 6.86 1.05 65 1 0.1032;

33.9 0.8518 7.987 -29 1.15 0.025 90 0.47 0.57 0.28 4.25 5.29 5.87 0.02 6.86 1.05 75 1 0.1472;

33.9 0.8518 7.987 -29 1.15 0.025 90 0.47 0.57 0.28 4.25 5.29 5.87 0.02 6.86 1.05 80 1 0.1685;

31 0.8664 6.535 -31 2.64 0.025 14.6 0.12 9.1 2 3.5 1.13 12.19 0.04 39.12 0.96 50 1 0.0668;

31 0.8664 6.535 -31 2.64 0.025 14.6 0.12 9.1 2 3.5 1.13 12.19 0.04 39.12 0.96 55 1 0.0812;

31 0.8664 6.535 -31 2.64 0.025 14.6 0.12 9.1 2 3.5 1.13 12.19 0.04 39.12 0.96 60 1 0.1038;

31 0.8664 6.535 -31 2.64 0.025 14.6 0.12 9.1 2 3.5 1.13 12.19 0.04 39.12 0.96 65 1 0.1373;

31 0.8664 6.535 -31 2.64 0.025 14.6 0.12 9.1 2 3.5 1.13 12.19 0.04 39.12 0.96 70 1 0.1496;

31 0.8664 6.535 -31 2.64 0.025 14.6 0.12 9.1 2 3.5 1.13 12.19 0.04 39.12 0.96 80 1 0.1704;

31.1 0.8666 12.72 22 3.11 0 3.29 0.19 6.1 1.5 4.5 2.35 8.5 0.05 41.3 0.82 50 1 0.0715;

31.1 0.8666 12.72 22 3.11 0 3.29 0.19 6.1 1.5 4.5 2.35 8.5 0.05 41.3 0.82 55 1 0.0932;

31.1 0.8666 12.72 22 3.11 0 3.29 0.19 6.1 1.5 4.5 2.35 8.5 0.05 41.3 0.82 60 1 0.1211;

31.1 0.8666 12.72 22 3.11 0 3.29 0.19 6.1 1.5 4.5 2.35 8.5 0.05 41.3 0.82 65 1 0.1478;

31.1 0.8666 12.72 22 3.11 0 3.29 0.19 6.1 1.5 4.5 2.35 8.5 0.05 41.3 0.82 70 1 0.1532;

31.1 0.8666 12.72 22 3.11 0 3.29 0.19 6.1 1.5 4.5 2.35 8.5 0.05 41.3 0.82 75 1 0.1677;

31.1 0.8666 12.72 22 3.11 0 3.29 0.19 6.1 1.5 4.5 2.35 8.5 0.05 41.3 0.82 80 1 0.1785;

32.6 0.8584 9.37 18 0.14 0 52.1 0.06 7.51 0.6 10.76 3.79 11.74 0.27 1.07 12.54 70 0 0.0414;

32.6 0.8584 9.37 18 0.14 0 52.1 0.06 7.51 0.6 10.76 3.79 11.74 0.27 1.07 12.54 70 2 0.0849;

32.6 0.8584 9.37 18 0.14 0 52.1 0.06 7.51 0.6 10.76 3.79 11.74 0.27 1.07 12.54 70 3 0.0624;

20.7 0.8894 38.56 -10 0.31 0.025 34.35 0.54 8.03 0 8.8 4.59 21.7 0.08 2.59 29.2 70 0 0.0403;

20.7 0.8894 38.56 -10 0.31 0.025 34.35 0.54 8.03 0 8.8 4.59 21.7 0.08 2.59 29.2 70 2 0.0815;

20.4 0.9276 43.06 -20 0.66 0.12 47.7 0.75 11.21 1.07 2.42 5.9 33.58 0.75 21.17 9.5 70 0 0.0651;

20.4 0.9276 43.06 -20 0.66 0.12 47.7 0.75 11.21 1.07 2.42 5.9 33.58 0.75 21.17 9.5 70 2 0.0978;

20.4 0.9276 43.06 -20 0.66 0.12 47.7 0.75 11.21 1.07 2.42 5.9 33.58 0.75 21.17 9.5 70 3 0.0863;

28.9 0.8782 13.65 13 1.12 0 11.5 0.38 0.16 1.06 8.67 3.12 70.73 0.07 46.64 1.53 70 0 0.0658;

28.9 0.8782 13.65 13 1.12 0 11.5 0.38 0.16 1.06 8.67 3.12 70.73 0.07 46.64 1.53 70 2 0.1042;

28.9 0.8782 13.65 13 1.12 0 11.5 0.38 0.16 1.06 8.67 3.12 70.73 0.07 46.64 1.53 70 3 0.0928;

33.9 0.8518 7.987 -29 1.15 0.025 90 0.47 0.57 0.28 4.25 5.29 5.87 0.02 6.86 1.05 70 2 0.1274;

33.9 0.8518 7.987 -29 1.15 0.025 90 0.47 0.57 0.28 4.25 5.29 5.87 0.02 6.86 1.05 70 3 0.1157;

31 0.8664 6.535 -31 2.64 0.025 14.6 0.12 9.1 2 3.5 1.13 12.19 0.04 39.12 0.96 70 0 0.0891;

31 0.8664 6.535 -31 2.64 0.025 14.6 0.12 9.1 2 3.5 1.13 12.19 0.04 39.12 0.96 70 3 0.1165;

31.1 0.8666 12.72 22 3.11 0 3.29 0.19 6.1 1.5 4.5 2.35 8.5 0.05 41.3 0.82 70 0 0.0976;

31.1 0.8666 12.72 22 3.11 0 3.29 0.19 6.1 1.5 4.5 2.35 8.5 0.05 41.3 0.82 70 2 0.1638;

]

m=n'

for i=1:19;

m(i,:)=2*(m(i,:)-min(m(i,:)))/(max(m(i,:))-min(m(i,:)))-1

end

m_date=m'

a=m_date(randperm(size(m_date,1)),:);%把样本按行随机打乱

a1=a(1:40,:);

a2=a(41:59,:);

p1=a1(:,1:18);

t1=a1(:,19);

p2=a2(:,1:18);

t2=a2(:,19);

p=p1';

t=t1';

%设置网络隐单元的神经元数

n=10;

%建立相应的BP网络

net=newff(minmax(p),[n,1],{'tansig','purelin','trainlm'});

%对没有训练的网络仿真

% y1=sim(net,p);

% hold on;

%figure;

%plot(y1,'b:+');

%title('样本仿真');

%xlabel('样本');

%ylabel('腐蚀率');

%keyboard

%对网络进行训练

net.trainParam.epochs=1000;

net.trainParam.lr=0.01;%学习率

net.trainParam.goal=0.00010;%目标误差

net=train(net,p,t);%训练

%对训练的网络仿真

% y2=sim(net,p);

%figure;

%plot(y2,'-+');

%title('样本仿真1');

%xlabel('样本');

%ylabel('腐蚀率');

%keyboard

%测试样本

tp=p2';

tt=t2';

ty=sim(net,tp);

%定义误差

tE=tt-ty;

tsse=sse(tE);

tmse=mse(tE);

%测试仿真

figure;

plot(tt,'-+');

hold on;%该指令是使后来的图形包含在同一个窗口中

plot(ty,'r:*');

legend('实验值','预测值');

title('BP网络模型输出预测曲线');

xlabel('输入样本点');

ylabel('腐蚀率');

axis([0,20,-1.5,1.5]);

tyr=(ty+1)*(max(m(19,:))-min(m(19,:)))/2+min(m(19,:));

我的网络，反归一化是不是不对呀。为什么归一后输出tyr 与归一化前ty一样呢？我感觉是(max(m(19,:)  min(m(19,:)

不对。因为他们输出分别为+1 和-1.

十分感谢！