Protecting assets through better data trending

Elaine Maslin examines how better data trending and a focus on key hubs could help improve ailing production efficiency rates and extend asset life on the UK Continental Shelf and beyond.

NI CompactRIO deployed on a vessel for measurement and control in a DNV certified application. Photo from  National Instrument.

Big data has been making headlines, not least in the oil industry. But, is big data—and how to handle it—being used to its full advantage across all areas of the business? Some think not, specifically with regards to maintaining and extending the life of aging production facilities on the UK Continental Shelf (UKCS).

Some 50% of North Sea facilities are at or beyond their expected design life, with many of those now expected to operate for even longer, as drilling, seismic and production technology extends field lives and recovery rates.

Operators have been focused on improving the integrity of these assets, but not all have fully embraced the next step—better aging and life extension (ALE) management, particularly through use of data collection and trending, according to the UK’s Health and Safety Executive (HSE).

ALE is a core focus under HSE’s Key Program 4, which follows the asset integrity-focused KP3. In a report, concluding its three-year investigation, HSE suggests better data and data trending could help to predict potential future failures and help plan maintenance.

The investigation looked at 33 installations on- and offshore on the UKCS. The final report alludes to insufficient key performance indicators, missing data and engineering drawings in places, and a tendency for audits to follow what happened prior, rather than addressing future needs. Referring specifically to electrical, control and instrumentation (EC&I) systems, the report cited a “widespread fix-on-fail approach.”

Such an approach could lead to down time—a sore point on the UKCS. The UK’s Department of Energy & Climate Change revealed production efficiency on the UKCS fell from 81% in 2004 to 60% in 2012. The poor efficiency levels are largely (nearly half) due to plant equipment failure and unplanned shutdowns, according to data from the Production Efficiency Task Force (PETF), set up by government-industry body PILOT. Worse, the basin is highly reliant on hubs. If they shutdown, it could impact multiple fields (see panel). Sir Ian Wood’s Maximising Recovery report, published earlier this year, also recognizes the need for better asset management, stating a “the need for significantly improving asset stewardship.”

Asset production efficiency trends in UK and Norway. Graph from McKinsey & Company.

A change in mindset is needed, says Tony Hetherington, head of operations, gas and pipelines, at the HSE. While ALE concepts are starting to be recognized and adopted, binary decisions are still being used to manage facilities, despite availability of systems for monitoring and trending data, he says.

“Binary decisions, i.e. assessing a piece of equipment based purely on it meeting its safe operating criteria or not, are no longer appropriate,” he says. “Computerized maintenance management systems will give them [operators or duty holders] the data they need, but they are not being used as much as they could be in a structured way across industry, because leadership do not see it as an essential thing to do.”

As an example, Hetherington cites an area high up the HSE’s importance list; safety critical elements (SCEs), such as emergency shutdown valves (ESDVs). An ESCV will be periodically tested, and if it meets performance criteria, it will continue to be used. Under a binary decision making system, its performance, or the time taken to close correctly, could have lengthened from the 8secs it was designed to close at, to 12 and then 18secs, but still pass its performance acceptability criteria, a phenomena described as “normalization of deviance” by the KP4 report. Another example is water pressure in a dousing system. If the pressure is falling year on year, it could be a sign it is silting up, but it could still pass the binary decision test, Hetherington says.

“If we trend that data now, it will give duty holders time to think about planning to do something about it,” he says, enabling better longer term asset management. “What is required now is looking at how equipment is performing over time, and what other parameters are changing, such as production fluid content, to assess how quickly equipment or pipelines are potentially degrading, and so when they might fail, and what the best maintenance or inspection regime might therefore be, and what action can be taken before it fails.”

Another example is using data trending to improve structural management, by helping to understand corrosion and fatigue. “If you can lose 50% of your wall thickness through corrosion, and the structure is corroding 2% per year, you know approximately how long you have got to use that facility,” Hetherington says. “By doing something about that corrosion rate now, maybe painting it more frequently, you could prolong the life of the asset.”

In practice

An NOV Hex mud pump. Image from National Instruments.

For fire and explosion systems, duty holders have good planned maintenance routines, using data trending, and there are increased inspection and maintenance routines on aging installations, including investment in replacing equipment and work on SCEs, with in-house core integrity teams and “system custodians” for specific SCEs, HSE says.

For structural analysis, concise structural reports are being created from inspection data, with anomaly trend analysis, and the findings reported to senior management (although, for some, not all structural analyses were up to date and some fatigue and redundancy analyses were incomplete or missing), the KP4 report says.

Risk-based assessments were also being performed, but these could be better utilized by looking at near-, mid- and long-term risks, such as managing the effects of changes in fluid properties and quantities, pressures, and souring.

In pipeline management, cross-industry collaboration has improved flexible riser inspection and integrity management, KP4 says. But, it continues, sophisticated corrosion modelling programs need to be validated, at suitable intervals and corrosion threat assessments, could be supplemented by longer-term corrosion mechanism predictions and management, “probably requiring greater integration of predictive information between reservoir and topsides engineers.”

Diagnostics

Using computer-based condition monitoring on assets is gaining more interest in the offshore industry, says Tristan Jones, Regional Marketing Engineer, Industrial and Embedded Systems, National Instruments, which supplies software and hardware used by engineers to develop systems that require measurement and control.

Jones says other industries are using more sophisticated data management and monitoring systems to manage aging assets, such as the railways and power generation sectors, by incorporating diagnostic systems. US-based Duke Energy, for example, has around 80 power plants and faces significant challenges with downtime resulting from component failure in older assets. However, for their predictive maintenance, they were spending 80% of their time taking measurements, and just 20% on analytics. They developed a more automated approach, by combining existing data from programmable logic controllers and sensors installed by original equipment manufacturers, with a significant installation of additional measurement nodes to evaluate the health of each asset. The data is aggregated to an enterprise level, with greater visibility, and the engineers can spend 80% of their time on prognosis, and predicting future failures, Jones says.

Another benefit of using automated, computer-based condition monitoring, incorporating data captured from PLCs or sensors installed by OEMS, and then using diagnostic systems, is that expert knowledge can be “harvested” and maintained in the system, and not lost when an expert on a particular piece of machinery leaves the business, Jones says.

Inefficient indirect hubs. Graph from McKinsey & Company.

Equipment uptime can also be increased, without having to send staff onto the drill floor, using automated sensor and diagnostic systems, Jones says. NI supplied a system to NOV for Hex mud pumps for use offshore in Norway. NOV wanted a computer-based condition monitoring system to maintain productivity on the pumps by allowing preventative maintenance, and without needing a human to access the pumps. A number of accelerometers, speed and phase sensors were added to the pumps to monitor vibration, in order to detect valve leaks. An NI CompactRIO system was added to the pump control system to acquire high-frequency data and power the sensors, along with signal processing software and alarm logics, using NI’s LabVIEW software. These enabled the team to monitor the pumps’ performance and schedule maintenance appropriately.

Different levels of prognostic tools can be added to monitoring systems, depending on requirements. Nick Ward, Senior Product Manager, Predictive Equipment Health Management, Controls and Data Services, at Rolls-Royce, has been looking at higher level prognostic tools, and sets them out into three generations. The first is threshold limiting. If a vibration value of a turbine, for example, goes over a specified limit, an alert is logged, setting reaction time, which, if breached, results in a shutdown. Second generation analysis involves an understanding of what the asset is supposed to do and how it should behave, by assessing live measurement data against stored data (showing what it should be doing) and finding anomalies, before raising an alert. Third generation analysis is a combination of second generation analysis, but incorporating additional factors, for example looking at vibration alone might not trigger an alert, but measurements on vibration, temperature and other parameters might, indicate a failure.

The tools for better ALE management are available, Hetherington says. “The industry, in general, hasn’t been taking advantage of the data it collects, particularly on aging assets. The industry is beginning to respond, but we need to keep up the pressure,” he says.

Strategic hubs

Tristan Jones, National Instruments.

Managing assets becomes even more critical when it comes to “strategic hubs” in the North Sea, according to McKinsey and Company, which presented a report on production efficiency at this year’s DEVEX conference in Aberdeen.

The McKinsey report found that the two main factors impacting performance are export system dependency and the quality of operator practices and approaches. McKinsey looked at production efficiency correlated with reliability practices in 2005/06 and then again in 2010/11. Best or good practice operators had higher asset production efficiency regardless of the year of study with leading performers also improving in the time period. Through interviews with 50 offshore field managers, McKinsey found that high performing operators were more likely to:

• Challenge and minimize planned downtime—by doing only the most essential activities during turnarounds and post-postponing others until normal operations resume, instead of treating a turnaround as a chance to tackle a “laundry list” of maintenance tasks that has been building up.

• Continually improve reliability by learning from failures—by routinely prioritizing and conducting investigations into root cause failures that result in lost production, then making appropriate changes to equipment, protocols or maintenance strategies.

• Create a culture of responsibility in operations—by instilling a sense of ownership in an asset, supported by widespread adoption of clearly visible and understandable operating standards.

Another key factor is reliance on hubs. Since 2000, indirect hubs and independents (i.e. fields that are tied to and rely on third party hubs for export), have had consistently lower asset production efficiency than direct hubs, and the performance differential is widening, McKinsey says. An outage at an installation housing a major export route has a severe knock-on effect on its dependents. “For instance, when a hydrocarbon leak was discovered at the Cormorant Alpha platform in early 2013, its operator closed the Brent crude oil pipeline which flows through it. The outage affected 27 other fields.”

Other recommendations from the report:

• Fix the basics of reliability and maintenance—employing strong operating practices, as well as having clear performance goals, rigorous root cause analysis, staff improvement schemes, and having a “one team” culture. Investing in facility upgrades and integrity-enhancing projects may be necessary as well.

• Establish standards in common operating tasks—including between operators and service companies, with a focus on knowledge sharing in areas such as common reporting and operating standards related to operating practices.

• Regenerate critical infrastructure within prominent hubs.

“Given the interconnectedness of the infrastructure we believe that the industry must develop a new operating norm at this hub level. First, production efficiency audits by a new regulator should begin by targeting operators of the most critical pieces of infrastructure,” McKinsey says. “The industry should also consider a more collective approach to managing ‘system critical’ infrastructure.”