If I have understood the question well, you are looking for a combinatorial interpretation of nth-fold self-convolution of a sequence. A quick google serach gave me this result on Catalan sequence. May be you can use similar idea for generic sequences.
Uday
- 2.2k
- 22
- 27