Page 49: of Offshore Engineer Magazine (Sep/Oct 2019)

48 50

REPORT Predictive Maintenance

Using live and contextualized data from the Ivar

Aasen platform, Framo is able to predict the status of its equipment and plan maintenance in advance.

Source: Aker BP in advance, Trond Petter Abrahamsen, completely new for us and has required guage processing and machine vision to

Managing Director of Framo Services the design of new smart contracts with continuously aggregate data plant-wide

AS, said in a statement when the “data Aker BP,” Abrahamsen said. sensor networks, enterprise systems, liberation contract” with Cognite and maintenance notes, and piping and in-

Aker BP was announced. “The new sys- strumentation schematics. Using histor-

An AI app tem sends intelligent data on our pumps, Oil? eld services company Baker ical and real-time data from entire sys- so we can predict how the pumps will Hughes, a GE company (BHGE) is also tems, AI and machine learning models perform in the future,” he said. advancing its predictive maintenance identify anomalous conditions that lead

Another goal of the inititaive is to capabilities. Its joint venture with en- to equipment failure and process upsets change the traditional approach to terprise arti? cial intelligence (AI) soft- across any number of assets and pro- maintenance, Framo said. Continuously ware provider C3.ai recently launched cesses, and then alerts the user to take ? owing live data allows for onshore BHC3 Reliability, an AI application that proactive action.

monitoring of equipment replacing un- A BHGE spokesperson told Offshore provides early warning of production necessary scheduled maintenance ac- downtime and process risk and gives Engineer that the company is currently tivities with maintenance when needed. failure prevention recommendations working with a customer on the imple- “While our service agreements previ- mentation of BHC3 Reliability and that and prescriptive actions.

ously just de? ned hourly rates, we will The generally available app uses deep a number of other customers have ex- now focus on uptime. This is something learning predictive models, natural lan- pressed interest.

SEPTEMBER/OCTOBER 2019 OFFSHORE ENGINEER 49