Aspen Plus DataFit中Std-Dev设置问题

longkui1990

原题见孙兰义教程例9.7第175页图9-66



第一行Std-Dev数据在没有输入之前默认值是两个1

后面三行数据是题目中提供的

这样运行的话会出现错误，错误内容同图9-71

参照官方bkp文件后，把Std-Dev数据换成两个0.01后顺利收敛

想请问下为何会出现这种问题？

Std-Dev有何作用？

如何确定这个Std-Dev的值？

谢谢！

烟台苹果

我也不懂 楼下解答吧

tdl522

Standard Deviation
The standard deviation is the level of uncertainty in the measurement. You can
enter it as an absolute or percent error (append a percent sign (%) to the value).
Statistically determined standard deviations are seldom available. It is enough to
supply an approximate “expected error,” estimated from experiences or
instrument specifications. Each residual (measurement — model prediction)
term in the sum of squares function is weighted by 1/(standard deviation).
You must specify a standard deviation greater than zero for each results variable
to be fit. If a zero value is entered for standard deviation, that results variable is
not included in the regression.
For inputs, a standard deviation greater than zero invokes maximum likelihood
(errors-in-variables) estimation: