SQL Server 2012 :: Pattern Matching (positive Across Records - Negative Within A Field)

I am doing some analysis on our customer base and their payment profiles. I have generated two profile strings, one for whether the balance of an account has gone up or down and one for the size of the balance in relation to the normal invoice amount for the customer. So (for example) the balance movement string will look like this:

UUUDUUUDUUUD-D00 Where U = Up, D = Down, - = no change and 0 = no change and no balance

I want to analyse these strings in two ways. The first is that I want to find customers with a similar pattern: in the example below the first and last patterns are the same, just one out of sync but should be considered the same

Movement Multiple CountRecords
UUUDUUUDUUUD1230123012301175
------------0000000000001163
UDUUUDUUUDUU3012301230121082

The second type of analysis is to find customers whose pattern has changed: in the examples above the patterns are repeated and therefore 'normal' in the records below the patterns have changed in that the first part does not match the second part.

Movement Multiple CountRecords
UUDUUUDUUUUU-----------07
UDUUUDUUUUUU------------7

good way to approach this without either a cursor or a hidden REBAR. The challenge as I see it is that I have to interrogate every string to find out if there is a repeating pattern and if so where it starts and how long it is (heuristic because some strings will start with a repeating pattern and then the pattern may change or deteriorate) and then compare the string for N groups of repeating characters to see if and when it changes and I can't think of an efficient method to do this in SQL because it is not a set based operation.