Different result command window and guide

Hello,

if I write in command window:

t=1
4.5582758767110180218764941687633*sin(314.0*t) - 16.557811525543529374688574837543*cos(314.0*t) + (16.557811525543529374688574837543*( 3.4028e+308*(1375.8856146096578468897575988832*t) + 0.48398685287825251446260421986529* 3.4028e+308*(1375.8856146096578468897575988832*t)))/exp(1485.7142857142857142857142857143*t)

I get NaN,but in guide I get -17.0711.....why?

Answer 1

[Mischa Kim]

John, in your expression you have two huge terms, basically,

 Inf/Inf

which results in a NaN. You should get the same result for the same expression in your GUI.

[john]

sorry....in guide I calculate it with eval:

t=0:0.01:1

result=eval(vpa((4.5582758767110180218764941687633*sin(314.0*t) - 16.557811525543529374688574837543*cos(314.0*t) + (16.557811525543529374688574837543*( 3.4028e+308*(1375.8856146096578468897575988832*t) + 0.48398685287825251446260421986529* 3.4028e+308*(1375.8856146096578468897575988832*t)))/exp(1485.7142857142857142857142857143*t))))

I got:

Warning: Rank deficient, rank = 0,  tol = NaN.

ggg =

Columns 1 through 4

 -16.557811525543528  16.565050277096358 -16.572247010655047  16.579401707964752

Columns 5 through 8

 -16.586514350877263  16.593584921351038 -16.600613401451270  16.607599773349886

Columns 9 through 12

 -16.614544019325649  16.621446121764180 -16.628306063157943  16.635123826106454

Columns 13 through 16....

and so on....

So what can I do with cosh, if cosh has big value? regexprep(result,'cosh',' 3.4028e+10*')? or regexprep(result,'cosh','0*')?

[john]

Or how can I plot for any time:

92.9*cos(314.0*t) + 96.3*sin(314.0*t) - (7.94*(cosh(1399.0*t) + 4.06*sinh(1399.0*t)))/exp(1500.0*t)

[Mischa Kim]

You can do

 syms t
 expr = 92.9*cos(314.0*t) + 96.3*sin(314.0*t) - (7.94*(cosh(1399.0*t) +  4.06*sinh(1399.0*t)))/exp(1500.0*t);

 expr = subs(expr, sinh(1399.0*t),(exp(1399.0*t) - exp(-1399.0*t))/2);
 expr = subs(expr, cosh(1399.0*t),(exp(1399.0*t) + exp(-1399.0*t))/2);
 expr = expand(expr,'ArithmeticOnly',true);
 ezplot(expr,[0,0.1])

The two subs titutions are hyperbolic trig identities, thereby avoiding the Inf/Inf issue (in combination with the expand command).