catanary arch problem -kangaroo

hi,

i am a civil engineer and i am playing around with kangaroo and gh. first i want to express my respect for this tools which are very straightforward and perceptive.

i modeled a little problem which is attached below. it`s an arch, let`s say out of bricks with loads, represented with point loads which are taken from user-defined surfaces (to represent self-wight e.g).

goal of the study is to find a supporting arch which is inside the arch then (provided the strength of the bricks is high enough) the arch is ok -otherwise one have to change the geometry of the arch (make it thicker, or change the rise of the crown).

therefore i used kangaroo in combination with galapagos to find a catenary which fulfill the boundary conditions. it works very well. so the solution galapagos found is very satisfying.

it`s simple to prove the resulting forces in the arch if one knows the rise of the arch.

according to the formula N~ q*l^2/(8*f)

with q= 21 kn/m, l= 10.00 m and f= 0.965

one get N= 272 kn, which is very near on the solution which kangaroo founds in the middle of the arch (271 kn). due to the point loads this force have to get slightly higher the nearer one comes to the anchor points. this works perfect too.

but there is one irritating thing.

at the end of the catenary, near the anchor points,  kangaroo gets two very different arch-forces. 281 kn in the next to last part. and 581 kn in the very last part. this is not possible and i am sure that the value of 581 kn is wrong. i calculated the example with a commercial fea-program too. it validates the kangaroo results except the first and the last one.

i think there is a problem with the calculated end length of the first (and last) element. they are twice too long as they have to be. or do i something wrong?

thank you for any reply and again for your work.

best peter