Intra_R_Add_Endocyclic rates are rubbish

[rwest]

There are only 5 rate expressions for Intra_R_Add_Endocyclic which then get averaged.
(See http://rmg.mit.edu/kinetics/Intra_R_Add_Endocyclic/ )

For example, Oj-CH=O closing to form a ring O1-CH-O-1 the forward rate I get is

OjCHO(15)(+m)=CHO2J(284)(+m) 1.0E0 0.0 0.0 !NetReaction from PDepNetwork #451 (HOCO) High-P Limit: Intra_R_Add_Endocyclic estimate: (Average of: (Average of: (Average of: (Average of: (R3_D doublebond_intra_pri_HDe radadd_intra_cs2H)) && Average of: (R3_T triplebond_intra_H radadd_intra_cs2H) && Average of: (Average of: (R5_SS_D doublebond_intra_pri_2H radadd_intra_cs2H) && Average of: (R5_SD_D doublebond_intra_pri_2H radadd_intra_cs2H)) && Average of: (Average of: (Average of: (Average of: (R6_SSS_D doublebond_intra_pri_2H radadd_intra_cs2H))))))) R3_CO carbonyl_intra_H radadd_intra_O
TCHEB / 290.0 2100.0 / PCHEB / 0.4934616333580064 98.69232667160128 /
CHEB / 5 3 /
CHEB / -2.0737000e+01 1.1524000e+00 9.6253000e-04 /
CHEB / 1.2862000e+01 3.3351000e-03 1.7455000e-03 /
CHEB / -2.9920000e-01 2.4939000e-03 1.3048000e-03 /
CHEB / -1.3781000e-01 1.5439000e-03 8.0725000e-04 /
CHEB / -5.8163000e-02 7.9064000e-04 4.1282000e-04 /

Because this 3-membered ring is very strained, the estimated thermo makes the reaction very endothermic, so the estimated reverse reaction rate is insanely high.

This family needs some work.

[rwest]

The automatic barrier-raising code merged by commit 072dbe8 should help a lot with this issue: endothermic reactions will have a barrier raised to be at least equal to the endothermicity.
Coupled with greg's QM calculations for cyclic species, and this shouldn't be so bad.

However, the fact remains that this family has very few data and thus produces very poor estimates.

I'm therefore leaving this issue open, but labeling it as "won't fix for 3.3"