Provide a workaround for the bug connected to calcualting confidence interval

[kduretec]

Fix or work around a bug in R that lets the confidence interval function crash when lower bounds are given to the parameter estimatoin method.
https://bugs.r-project.org/bugzilla3/show_bug.cgi?id=15960

[kduretec]

It seems that this bug is caused by the nls and nlsLM implementations.
casallas/minpack.lm#1
wch/r-source@0b6a9fc

[kduretec]

The workaround will be not to use lower bounds in the model estimation and discard all estimated moels whose paramters are smaller than 0.