Offshore Logistics: Forecasting Product to Predict and Optimize Repair Routes

A large National Oil Company (NOC) spent more than $300 million annually on offshore-related logistics. Logistics planning was performed manually based on limited information, resulting in inconsistent asset utilization rates and sub-optimal routes and vessel chartering. Supply chain decisions were reactive, primarily driven from asset events after they occurred. For example, a recent equipment failure required replacement components to be shipped from an overseas supplier due to lack of inventory from local suppliers. These inefficiencies were a key driver of sub-par financial performance at the rig level, where one unit reported more than 1,400 hours of non-productive time (NPT) in the year related to logistic delays.

Arundo’s data scientists worked with the company’s internal analytics team to build a machine-learning model to predict logistics needs based on equipment operational data combined with forward-looking drilling plans. This model was then deployed and operationalized within the Arundo platform. All within 120 days.

The solution was delivered in two stages:

1. Built a predictive model utilizing historical information to map and analyze key operating variables.
2. Integrate and deliver a holistic solution that allowed management to plan supply vessels based on demand.

RESULTS

By using the Arundo suite, the custom predictive solution was built and deployed at the customer site within 120 days. Based on model accuracy and historical performance, the customer is expected to see an NPT decrease of 15% and an increase in equipment utilization of 22% in 2017.