Knowing the future: predictive maintenance and risk mitigation
The advent of monitoring and predictive technology has to this day averted a myriad of accidents and millions of dollars in equipment failures and plant downtime. At a time when asset optimization is the word of order and when security and environmental concerns are higher than ever before, predictive maintenance can prove to be the answer to the prayers of managers, health and safety executive directors, and engineers alike.
"The biggest challenge in the oil and gas industry today is the identification and forecasting of hidden risk development," said Henk Smith, director of key accounts for Europe and South America at GP Rovsing Dynamics, a GP Strategies Company.
When to implement predictive maintenance technology
"In our experience, operators that engage with predictive maintenance technology are astonished at the results, but the principle strategy we apply is in fact quite basic. It is just a pragmatic approach to problem solving with technology that has been available for quite some time but [one] that companies have so far been slow to implement," Smith added.
This disparity between available technology and lagging industrial application can sometimes be perplexing.
Let’s consider this: As the manager in charge of maintenance and reliability at an industrial plant, between the choice of running your equipment to exhaustion and then replacing it (reactive maintenance), replacing the equipment periodically according to manufacturer instructions whether it is damaged or not (preventive maintenance), or having a system that would tell you in advance what kind of damage your equipment is suffering and when is the optimal time to replace it (predictive maintenance), which option would you choose? It depends.
For cheap, noncritical and easy-to-replace equipment, installing high-tech sensors that flood your servers with performance indicators might be unjustifiably expensive and useless. It is just easier to replace it once it stops working. Noncritical equipment that has a long lifecycle that is subjected to little stress and degradation might also not justify constant monitoring if it were replaced periodically. However, if you could predict when the moving parts of a gas processor would start to degrade and have the ability to act on it at an optimal time to reduce costs and prevent catastrophic failure, then your investment might be well worth it.
A good example of implementing predictive maintenance technology was a near catastrophic event at Total’s Petrochemical facility in Carville, La. Using infrared thermography, Azima DLI, a provider of predictive condition monitoring and analysis services who had been working at Total Petrochemical’s plant applying regular predictive maintenance analysis since 1998, was able to detect an anomaly in an aluminum bus bar inside the plant’s reactor. The analysis revealed a temperature measurement at the jumper wire connection with the reactor above 800 ° F, which would had led to eventual shutdown and incalculable destruction of surrounding equipment and safety risks for workers. An inadequate clamp in the jumper wire was identified, saving Total Petrochemical thousands in repairs and reduced downtime.
Implementing predictive maintenance technology
Timing is key when applying predictive maintenance systems and naturally, the earlier the better. "Reliability is crucially important and we see more and more oil and gas companies paying attention to these issues, not only from an economic point of view, to avoid critical failure, operational shutdown, and equipment damage, but also from a health, security, and environmental protection point of view. Above all, in order to take all the benefits from these technologies, it is fundamental that they are applied as early as possible in developing projects, as application is costly and more complicated when operations are already on the way. There is a great advantage to integrating these technologies in the front-end engineering design (FEED) process," noted Paul Wheelhouse, director at Red Wheel Solutions, a maintenance, asset management, and reliability consulting company.
Another project developed by Azima DLI included the ConocoPhillips refinery in Ferndale, Wa. Built in 1954, the aging refinery was an integrated single train fuels refinery with FCC, HF alkylation, and reformer operating units, with a processing capacity of 110,000 barrels per day. The plant had a reactive maintenance approach with staff constantly concentrating on failing pumps and motors, which prevented it from having any long-term maintenance strategy. ConocoPhillips chose to install vibration monitoring sensors and use integrated analysis and information-processing software, which uses statistical analysis to compare accumulated average baseline signatures from the pumps against the test data. By comparing the tested machine’s spectral peaks to this baseline, excesses are processed through a set of rules for the machine, and a diagnosis is rendered. According to Azima DLI, the plant’s pumps’ rate of mean time between failures fell from 22 months to 52 months in a period of 10 years, and maintenance costs for the refinery’s pumps dropped by 50%. The ConocoPhillips refinery in Ferndale now has a 91% plant availability rate.
Benefits of predictive maintenance technology
Industry players suggest that the average period of return on investment for these systems is 12 months, with escalating financial benefits over time. That seems, however, not always to be the main concern for operators.
"In oil and gas, cost is not always the primary objective. We have come to understand that the most recognizable advantage of these technologies for oil and gas companies is focused on risk mitigation rather than return on investment," Burt Hurlock, CEO at Azima DLI said.
Through analysis and cross industry comparison, today these companies are capable of developing:
- Risk analysis
- Aggregation of all mechanical faults detections in operations
- Scored mechanical fault detections by severity and potential impact for disrupting production and operational continuity
- An overarching view of the system’s potential benefits for risk aversion.
As concerns for human security rise in oil and gas operations, these approaches can be instrumental in saving lives and promoting operational uptime.
Improving predictive maintenance practices
A presentation at Power-Gen International conference in 2014, made by Helmer Andersen, performance and condition monitoring manager at GP Strategies Corporation, and Bruno Lima, technical communications PL from Petrobras, described the application of real-time fleet condition and performance monitoring to Petrobras’ gas-fired power plants in Brazil to improve operational availability and reduce fuel and maintenance costs. Seventeen power plants, with a combined output of 6 GW, are now monitored across the country. The project includes the tracking of 63 gas turbines, 11 steam turbines, 59 gas engines, and 27 heat recovery steam generators through advanced performance, combustion, and advanced condition monitoring using thermal modeling, vibration spectral monitoring, and reliability assessment. The operation is managed using software at the monitoring and diagnostics center, located in Rio de Janeiro. Initially implemented in 2013, phase 1 of the project lasted for 6 months including IT infrastructure, import of 24.000 parameters, with basic performance and condition monitoring. Phase 2 took 12 months to implement advanced condition monitoring using spectral vibration monitoring for critical production machinery, and further implementation of the advanced functionality is still on-going.
"The monitoring approach with full automation of diagnostic data evaluation, fault diagnosis, and fault forecasting revealed major problems in several plants that could be addressed before a critical situation would be reached, avoiding unscheduled downtime. This changed the operation from reactive to proactive, the key objective of the project, and realized a short return on investment," Smith said.
An older, but demonstrative case, was also detailed in a paper authored by Smith and TAQA Energy’s Neil Macrae regarding the application of predictive maintenance for the gas compressor trains of the North Sea production platform P15D, owned by BP Netherlands Energy’s (today the facility is owned by TAQA Energy), in 2003. In this case, the challenge for BP was "operating a mature gas field with a constant need for cost reduction while maintaining availability on-demand." An intelligent condition, performance, and reliability monitoring system (CPRM) was put in place to quantify historic and future risk of downtime and assess capacity reduction to manage operations and maintenance. A dedicated reliability monitoring software package was used to identify the main causes of downtime over the previous 4 years.
The objectives of the intervention included:
- Online condition monitoring covering both the gas turbines
- Gearboxes and compressors
- Online performance monitoring covering the gas turbines and compressors
- Online reliability monitoring covering the compressor lines and 30 less critical machines
- Offline condition monitoring of electric motor driven fans and pumps.
The system was designed so that all information would be processed through one single server at the platform that would allow for onshore monitoring and management from any one of BP’s global offices using particular software. According to the previously mentioned study, in 2 years, reliability was improved by 1%, from 96% to 97%, and availability improved by 2%, from 96% to 98%. The system, including implementation, had paid for itself in 12 months. One of the most relevant features in this particular application is that the CPRM system distinguishes between information for operators, maintenance engineers, process engineers, and management with each user group having its own warning system with dedicated information related to its responsibility. Operators receive immediate information about actions that need to be executed in order to minimize and shorten trips, maintenance engineers receive information about long-term problems, process engineers get information about deteriorated performance, and managers get information about achieved key performance indicators on reliability, availability, and utilization, plus variations in the benchmark figures.
These technologies have been applied for a very long time, but the information went to plant managers and technicians alone. What has changed in recent years is the advent of the Internet, which has finally allowed aggregating that data across a number of production and operation plants and applying it vertically to any company. "What the digitalization has allowed providers is that we are now able to take business level metrics and benchmark the performance of our customers. This allows them to rank the performance of their different operations, to identify which ones run well and which ones don’t, and also to benchmark their performance against other companies and operators in the market," Hurlock said.
The most common uses of predictive maintenance technology, including oil analysis, infrared thermography, and vibration analysis, have been around for over 3 decades, but have failed to attract the attention of managers that tended to see maintenance issues as minor, technical affairs. The industry still finds it hard today to send the correct message to decision makers in operating companies, but that might be changing.
Obstacles to using a predictive maintenance program
Across the board, industry players pin down poor internal communication structures and old-fashioned corporate culture as the reason for these technologies to have lagged in application. The cost didn’t seem relevant enough for a manager to care for it at all. However, once it became known that these technologies had a considerable physical-risk mitigation aspect, the attitudes slowly started to change.
"Communication between technicians and management is always an issue in many organizations, and that is the case in the application of these technologies as well," Wheelhouse said. "The major obstacle to optimization is the difficulties of communication between the plant floor and the decision-making board. That will not change quickly. In our experience, the flow of communication between the technical side of the operation and management is very dependent on the company’s culture. If the company upholds an old-fashioned culture, they will be very protective of the information they share even amongst themselves, to the extent that the technicians will not want management to see how well or how poorly they are operating. They are asked to do a lot without very much. They will be asked to implement a strategic predictive maintenance program without enough funds to do it and with no specific strategic objectives," Hurlock said.
The ability to benchmark and compare performance as well as a risk-focused promotion of these technologies should now allow for managers to see the risk mitigating and strategic importance of these technologies; however, another problem subsides. Investment in maintenance procedures doesn’t tend to be too high up on oil and gas companies’ agenda, and a sort of Catch 22 emerges. Basically, "when the oil price is up, nobody spends money on maintenance because they are making so much money they don’t have to, and when the prices are down, there is a very myopic view of cutting costs. It will take leadership and a culture change for these technologies to be implemented, and that will take a long time," Hurlock said.
In the meantime, the gap between maintenance performance between top optimized players and others will keep enlarging. As a 2014 study by Mckinsey & Company indicates, "the performance gap between industry leaders and other companies widened, from 22 percentage points in 2000 to around 40 percentage points in 2012." Performance benchmarking now allows us to compare the benefits and downfalls of systems and procedures alike. The costs of catastrophic events, as accidents like the Macondo disaster of 2010, keep reminding us, make a bigger impression than a simple economic impact. Loss of life and environmental contamination are unacceptable consequences of the search for oil, but only through embracing a risk mitigation approach can these events be minimized.
João Marques is a contributing writer to Oil & Gas Engineering.