Message from the Machine

1. Message from the Machine

2. Who are we ? • Worldwide leader in the design, manufacture and sale of equipment and components used in oil and gas drilling • The provision of oil field inspection and other services and supply chain integration services to the upstream oil and gas industry Established: 1841 NYSE: NOV Employees: 60,000 Revenue (2012): $20b

3. Not just the Oil Field… • We solve really big problems

4. Message from the machine…. So what does that mean ? • Condition Monitoring • Condition Based Maintenance • Predictive Maintenance • No way.. Not in the drilling industry !!!!

5. The challenges we face • Extremely Remote Locations • Extremely Rugged (Rough Neck Proof) • Low Connectivity • Extremely Expensive when things break • Antonov AN 124 rented to fly equipment • Technology Averse (unless it turns to the right !) • Tried and Failed a few times… No Trust it will work…

6. Why can’t we (Drilling Industry) do it ? • Automotive • 100 + prototypes, 30 cars at each pre-production • Airlines • 747 Example – Redundancy of Systems. Collaboration of partners • Production • Been doing it since the 80s, no moving assets, but similar asset classes • Aerospace (NASA) • Only one chance, it has to work – Testing simulation and software • This is the model NOV plans to replicate • – Data Analysis, Correlation Analysis

7. Why we should be doing it ? • Current Day Rate : > $500,000 • Typical Downtime on a fleet : 4% • Annual Downtime for a 10 Rig Fleet: $72,000,000 • If Downtime is reduced to 3.75% from 4%: Savings of $4,500,000 / yr. • That goes to the bottom line… • Or increase Revenue by $18,000,000 @ 25% OP. • Now imagine Spread Costs which is over twice those numbers

8. It’s not just the financial impact! • Incidents in North Sea, Alaska, Gulf of Mexico • We are all responsible

9. Maintenance Savings • Condition Based not Calendar Based • Extend Maintenance Schedules • Standard - 4 oil changes by 12000 miles • New – 1 Oil change at 10500 miles • Translate that to a fleet of 100 rigs x 3 engines or more

10. IADC/SPE 128865 - How do we do this ?

11. Criticality Top Drives & Blow Out Preventers • Specifications • Continuous Torque 105,000 ft.-lbs. • at 130 rpm • Horsepower 2 * 1340 HP • Max Speed 250 rpm • Break-Out Torque 150,000 ft.-lbs. • Static Load 1,250 Ton Hoisting • Drilling Load 1,226 Ton Rotating Torque Speed Bearing Life One company … unlimited solutions

12. Process • Analyzed Failure Data • Over 5 Large Offshore Drilling Contractors • Maintenance Data • Operational Data • Criticality of the failures • Probability of Failures • Consequence of Failures • Why can’t we stick a sensor on there ?

13. Top Drive Boundary Diagram

14. Detailed Breakdown of Failures

15. Data Analysis • Started with 10 years worth of data • Finally Ended up with 3 years worth of good data • “Other” – largest cause of downtime

16. Condition Monitoring System • Intelligent data acquisition and analysis Quality of prediction • Accounts for changing operating conditions • Accounts for random events such as jarring Filters out ambient noise and resets baseline

17. Vibration Sensors

18. A day in the life of …. Boiler • Gets to location • Stays there • +30 years

19. A day in the life of …. Top Drive

20. Advanced Rig Monitoring – Virtual Employee eHawk introduces Machine Learning to learn from historical data an equipment’s “Normal” vs. “Abnormal” operations for multiple states. Dynamically create profiles to monitor the equipment and listen for abnormal signatures. Continuously adapts over time.

21. The Message…. How do we do it ?

23. What do we measure on a Top Drive then Top Drive RPM • Upstream • Engines, Generators, Drives • Downstream • Rock formations, DH vibrations, etc. • Other • Ambient Temperature, etc. Torque Load Derrick Location Motor Winding Temps Left Right

24. Machine Learning

25. Example 1 - Top Drive Found 14 day Predictor for June10th Top Drive Failure (Over Temp) due to Lube Motor Failure Created Supervised Learning Profiles to monitor for failure signature detected ongoing. Confirmed failure information with Rig Root Cause: UV1 valve cycling lubrication fluid in 6 second intervals

26. Example 2 – Engine Temps

27. Example 3:– Mud Pump • Drive fault prediction found • Customer notified • Rig down 18 hours later Anomaly Detection triggered on June 9th for Mud Pump Drive fault

28. So... No sensors whatsoever ??? Cam and Cam Main Bearings

29. 3 Day Trend shows increased vibration. • The spectrum and trend of the BPFI (Ball Pass Frequency Inner Race) clearly indicated the presence of an early stage defect on the inner ring of the bearing Screen shot of Actual Alarm Frequency Spectrum 3 day trend of BPFI

30. Actual Failed Bearing

31. Associated Costs of Components and Labor for this failure Note: Spare Eccentric Cam was not available on Rig Lead time for Eccentric Cam was extremely long at time of incident

32. Risk Analysis

33. The Numbers Say it All Without solution $173,000 Lost Day Rate With solution (2012) $33,000 Lost Day Rate 72% Reduction in RigDown

34. Predictability • False Alarms • Predictability is key • When you call the rig with a problem…

35. Monitoring the BOP

36. Do we have to do this? • Current Demographic – • Older, aging driller demographic • Influx of younger inexperienced workforce • New Deepwater Rigs being built – How will we staff them • Compare Pilots and Drilling Crews ($650 Million Rig) • Not enough time to train – Serve the iPad Generation

37. Remote Monitoring – Too many rigs and too few PhDs Hawaii Norway Singapore Houston

38. Conclusion • Are we there yet ? • There is no end point. • We can only get better • Technologies and Capabilities exist • Let’s use the tons of data • We have no more excuses

39. Questions?

40. NOV I Rig Solutions I www.nov.com