Why do I get 3 principal stress when solving a Plane stress (C_PLAN) problem??

This is really extrange for me, but as far as I know if the problem is plane stress I should have only 2 principal stress (I'm using C_PLAN if this is the problem please let me know), but when I extract the results to a table I get this

[code:1]
#Reac/Force at nodes
#Principal Stress
,INTITULE ,NOEUD ,RESU ,NOM_CHAM ,NUME_ORDRE ,INST ,ABSC_CURV ,COOR_X ,COOR_Y ,COOR_Z ,DX ,DY ,VMIS ,TRESCA ,PRIN_1 ,PRIN_2 ,PRIN_3 ,VMIS_SG ,TRSIG
,ReacXYZ ,N2 ,RESU ,REAC_NODA , 1 , 0.00000E+00 , 0.00000E+00 , 5.91974E-17 , 1.25767E-17 , 0.00000E+00 , 4.32859E+07 , 2.40147E+07 ,- ,- ,- ,- ,- ,- ,-
,ReacXYZ ,N1 ,RESU ,REAC_NODA , 1 , 0.00000E+00 , 1.00000E+01 , 5.00000E+00 , 8.66025E+00 , 0.00000E+00 , 1.78576E+01 ,-4.80295E+07 ,- ,- ,- ,- ,- ,- ,-
,ReacXYZ ,N3 ,RESU ,REAC_NODA , 1 , 0.00000E+00 , 2.00000E+01 , 1.00000E+01 , 0.00000E+00 , 0.00000E+00 ,-4.32859E+07 , 2.40147E+07 ,- ,- ,- ,- ,- ,- ,-
,Esfuerzos ,N137 ,RESU ,EQUI_NOEU_SIGM , 1 , 0.00000E+00 , 0.00000E+00 , 7.99715E+00 , 1.31366E+00 , 0.00000E+00 ,- ,- , 2.14783E+07 , 2.15065E+07 ,-2.14720E+07 ,-2.20640E+04 , 3.45016E+04 ,-2.14783E+07 , 0.00000E+00
,Esfuerzos ,N246 ,RESU ,EQUI_NOEU_SIGM , 1 , 0.00000E+00 , 5.50244E+00 , 2.50229E+00 , 1.60226E+00 , 0.00000E+00 ,- ,- , 1.72916E+07 , 1.74341E+07 ,-1.74341E+07 ,-2.89568E+05 , 0.00000E+00 ,-1.72916E+07 , 0.00000E+00
,Esfuerzos ,N465 ,RESU ,EQUI_NOEU_SIGM , 1 , 0.00000E+00 , 8.88197E+00 , 3.67078E+00 , 4.77336E+00 , 0.00000E+00 ,- ,- , 1.16345E+07 , 1.18679E+07 ,-1.18679E+07 ,-4.86831E+05 , 0.00000E+00 ,-1.16345E+07 , 0.00000E+00
,Esfuerzos ,N498 ,RESU ,EQUI_NOEU_SIGM , 1 , 0.00000E+00 , 1.12077E+01 , 5.49592E+00 , 6.21482E+00 , 0.00000E+00 ,- ,- , 1.98208E+07 , 1.98574E+07 ,-1.98006E+07 ,-1.65611E+04 , 5.67891E+04 ,-1.98208E+07 , 0.00000E+00
,Esfuerzos ,N508 ,RESU ,EQUI_NOEU_SIGM , 1 , 0.00000E+00 , 1.45725E+01 , 5.00000E+00 , 2.88675E+00 , 0.00000E+00 ,- ,- , 7.02033E+06 , 7.06523E+06 ,-7.06523E+06 ,-6.97455E+06 , 0.00000E+00 ,-7.02033E+06 , 0.00000E+00

[/code:1]

If I see the model with Salome postprocessor, there are also 3 principal stress, so the problem is not when I extract the values.

The problem is a triangle in the xy plane (I try other geometry but the result is the same) with a displacement imposed at each corner (to simulate a compressive force)

[code:1]
CHAR=AFFE_CHAR_MECA(MODELE=MODE,
DDL_IMPO=(_F(GROUP_NO='p1',
DX=0.001,
DY=5e-4,),
_F(GROUP_NO='p3',
DX=-0.001,
DY=5e-4,),
_F(GROUP_NO='p2',
DX=0.0,
DY=-0.001,),),);
[/code:1]

[code:1]
MODE=AFFE_MODELE(MAILLAGE=MAIL,
AFFE=_F(TOUT='OUI',
PHENOMENE='MECANIQUE',
MODELISATION='C_PLAN',),);
[/code:1]

If anybody can tell why this happend I'll appreciate it, I'm trying to aproximate this values using another method but I get large errors in the points with 3 principal stresses

Well it seems like nobody knows why this happens, anybody can tell me how to write the the stress according to xyz directions?