同一个版本7.11.0.584 (R2010b)，安装在一个32位台式机上，和一个64位笔记本上，操作系统都是win7。

同一个.m文件，在32位台式机上运行的好好的，但是在64位笔记本上就出错了！报错如下：

??? Error using==> mtimes

Inner matrixdimensions must agree.

Error in ==>evalincaller at 14

r =evalin('caller',a);

Error in ==>sym.eval at 33

r =evalin('caller',['evalincaller(''' sym2str(a) ''')']);

Error in ==>ColdHRBWO7 at 247

plot(rx,eval(SS(kn,1))/TP,'k','LineWidth',sqrt(abs(kn)));

经过检查，问题出在：

SS(kn,1) =real(omg*epsl*T(kn,1)*T(kn,1)*bt(kn,1)*A(kn,1)*A(kn,1)*besselj(1,T(kn,1).*r).*besselj(1,T(kn,1).*r));

上。于是，我做了如下尝试：

(1)     我改成：

SS(kn,1) =real(omg*epsl*T(kn,1)*T(kn,1)*bt(kn,1)*A(kn,1)*A(kn,1)*besselj(1,T(kn,1).*r));

在64位笔记本上就不出错了。但真正的公式里面，的确需要两个bessel公式相乘。

(2)     我又把两个bessel之间的.*改成*：

SS(kn,1) =real(omg*epsl*T(kn,1)*T(kn,1)*bt(kn,1)*A(kn,1)*A(kn,1)*besselj(1,T(kn,1).*r)*besselj(1,T(kn,1).*r));

64位笔记本还出错！报的错误仍然是“Inner matrix dimensionsmust agree.”

(3)     我又改成如下两种形式：

SS(kn,1) =real(omg*epsl*T(kn,1)*T(kn,1)*bt(kn,1)*A(kn,1)*A(kn,1)*(besselj(1,T(kn,1).*r))^2);

SS(kn,1) =real(omg*epsl*T(kn,1)*T(kn,1)*bt(kn,1)*A(kn,1)*A(kn,1)*(besselj(1,T(kn,1).*r)).^2)

64位笔记本继续出错，报的错误又成了：“Inputs must be a scalarand a square matrix.”

出问题的程序段如下：

syms r z;

SS = zeros(2*n+1,1);

SS = sym(SS);

for kn=1:1:2*n+1

SS(kn,1) =real(omg*epsl*T(kn,1)*T(kn,1)*bt(kn,1)*A(kn,1)*A(kn,1)*besselj(1,T(kn,1).*r).*besselj(1,T(kn,1).*r));

end

%#########################谐波功率#########################;

power=zeros(2*n+1,1);

TP = 0;

for kn=1:1:2*n+1

clear sp;

sp = @(z,r)real(besselj(1,T(kn,1).*r).*besselj(1,T(kn,1).*r).*r);

power(kn,1) =real(2*pi/p*omg*epsl*T(kn,1)*T(kn,1)*bt(kn,1)*A(kn,1)*A(kn,1)*quad2d(sp,0,0.5*p,0,Rw,'RelTol',1e-5,'MaxFunEvals',9000));

TP = TP + power(kn,1);

end

power = power/TP;

save power;

xlswrite('5-谐波功率.xls',power);

%#########################功率密度#########################;

rx = 0:0.01:(ro-d)*1e3;

r = rx*1e-3;

z = 0.5*p;

GLMD = rx';

figure(100);

for kn=1:1:2*n+1

clear GLMDn;

plot(rx,eval(SS(kn,1))/TP,'k','LineWidth',sqrt(abs(kn)));

hold on;

grid on;

box on;

xlabel('r-axis (mm)');

ylabel('Power flux density (W/m2)');

GLMDn = eval(SS(kn,1))/TP;

GLMDn = GLMDn';

GLMD = [GLMD,GLMDn];

end

save GLMD;xlswrite('5-Poynting.xls',GLMD);

0