This is a gaps and islands problem, but I would use the difference of row numbers approach:
select orderid, count(*), min(serialNumber), max(serialNumber), type
from (select t.*,
row_number() over (partition by orderid order by serialnumber) as seqnum,
row_number() over (partition by orderid, type order by serialnumber) as seqnum_type
from t
) t
group by orderid, type, (seqnum - seqnum_type)
order by orderid, min(serialNumber);
Your serialnumber looks like a string, but it has numeric values. If you can trust the sequencing of it, you don't even need both row number values:
select orderid, count(*), min(serialNumber), max(serialNumber), type
from (select t.*,
row_number() over (partition by orderid, type order by serialnumber) as seqnum_type
from t
) t
group by orderid, type, (serialnumber - seqnum_type)
order by orderid, min(serialNumber);
The reason this works is a little hard to explain, but quite obvious if you run the subqueries. You will see how the difference between the row numbers identifies the groups that you want to identify.
Much cleaner ... I just got the Flag approach stuck in my little head +1