rbf.m

rbf神经网络在变压器故障诊断中的应用

p=[0.045036 0.108780 0.128584 0.009097 0.000000;
   0.119754 0.646052 0.279757 0.176645 0.030556;
   0.017400 0.009939 0.005213 0.023694 0.727778;
   0.148414 0.657096 0.384014 0.182991 0.000000;
   0.035824 0.267808 0.693310 0.025809 0.005556;
   0.078813 0.828272 0.357950 0.582822 1.000000;
   0.031218 0.020431 0.008688 0.029194 0.944444;
   0.058342 1.000000 1.000000 1.000000 0.111111;
   0.631525 0.309221 0.184188 0.053945 0.694444; 
   0.662231 0.276091 0.288445 0.041252 0.127778;
   0.160696 0.701270 0.279757 0.202031 0.000000;
   0.006141 0.018222 0.065161 0.007193 0.010000;
   1.000000 0.386527 0.496959 0.024328 0.022222;
   0.000000 0.675870 0.707211 0.818913 0.150000;
   0.018731 0.036996 0.004334 0.002962 0.000000]';
t=[ 1 0 0 0 0 ;
    0 0 0 0 1 ;
    0 0 1 0 0 ;
    0 0 0 1 0; 
    0 0 0 1 0; 
    0 0 0 0 1; 
    0 0 1 0 0; 
    0 0 0 0 1; 
    0 0 1 0 0;
    0 1 0 0 0; 
    0 0 0 1 0; 
    1 0 0 0 0; 
    0 1 0 0 0;
    0 0 0 0 1; 
    0 0 0 1 0]';
%建立并训练网络
for i=1:5
    net=newgrnn(p,t,i/10);
    y(i,:)=sim(net,p);
end
%绘制网络的逼近效果
plot(p,t);
hold on;
plot(p,y(1,:),'*');
hold on;
plot(p,y(2,:),'d');
hold on;
plot(p,y(3,:),'ro');
hold on;
plot(p,y(4,:),'kp');
hold on;
plot(p,y(5,:),'g+');
hold off;
figure
%绘制网络的最佳逼近误差
plot(p,y(1,:)-t,'+-');