Challenge

The offshore Rigs requires vessels for their supplies. Currently, based on the materials request raised, the number of vessels required are decided manually and on the fly. The challenge is to not only make this process automated but also predict the number of vessels that will be required in the future based on the activities going on in a rig to allow rig operators to plan and book the vessels beforehand.

Approach

Two fold approach was taken, based on the three major type of Operations which happens on the Rigs- Time series forecasting was used to forecast, how much Surface area(m2) and Volume(m3) will be required, these were in turn used to calculate number of vessels will be required in the future. Second Approach was use of ML algorithms for Drilling operations only. Here the number of vessels required was predicted directly. Triangulation of the two approaches was used to arrive at an accurate forecast.

Outcome

Accurate forecasts are expected to lead to a 3% reduction in number of vessels and a corresponding decrease in cost of operations.