My degree is in maths (with a specialism in statistics).
In your situation my instinct would be to calculate the total running costs over the past year (or at least over an extended period where the relevant factors are largely the same as they are now). Then, of course, divide by the number of working days and (if relevant) add on a small percentage to allow for inflationary changes from the mid-point of the assessed period up until now.
However I might then want to refine that model in order to take into account factors which might have changed significantly (i.e. by more than the rate of inflation) over the relevant period. For example if road fuel costs contribute significantly to the result then (since they've risen well above the rate of inflation) they would need to be weighted accordingly.
Chris