Data-Driven Maintenance Wins At Petronas
Grant Gerke | February 20, 2019
Digital initiatives are starting to mature within enterprises and success stories are making their way to company websites. Many of these successes are coming from the oil and gas industries. Overall, this sector has been one of the real winners with data-driven-maintenance projects, partly due to its need to innovate when oil prices plunged in the middle part of this decade.
Petronas Global, Kuala Lumpur, Malaysia (petronas.com), a major supplier of oil and gas, recently documented its move to data-driven maintenance routines for upstream activities, such as well and pipe activity in Malaysia. Enlisting a technology platform from OSIsoft, San Leandro, CA (osisoft.com), it first modeled essential rotating equipment for upstream operations using a mix of historians, modeling, and visualization tools.
In that first phase, Petronas rolled out its Protean solution (the company’s internal name for the pilot) by developing a monitoring system for just two critical gas-turbine-driven compressor units. It used OSIsoft’s Asset Framework (AF), part of the PI System, to build standard equipment templates to model all of its compressors, regardless of manufacturers. The AF technology provides the ability to contextualize and analyze data from multiple sources, including one or more of OSIsoft’s PI historians, internal lab databases, and external relational databases.
The heart of the project’s challenge was the company’s external databases. As Timur Tashkenbaev, a Petronas prodution specialist, noted in his presentation at the 2018 PI World Conference, “We were having daily issues with disparate databases, and generating a report could take more than seven days.” On certain days, he said, an oil well would quit, and it would take three days from detection to react to it. According to the company, a production day for one well can be in the range of 30,000 barrels.
The AF module ties into other OSIsoft systems, such as PI Vision and PI ProcessBook, to build operational displays, offer calculations, and deliver critical alarms. According to Tashkenbaev, Petronas built enhanced dashboards, established alarm systems, and started to embed workflows into systems in phase one. “And, now,” he explained, “we have deployed phase two and have a full set of the workflows connected to our corporate databases, such as well test data and sampling data.”
By mid-2017, Petronas had realized more than a million dollars in savings through predictive maintenance and was confident enough to push the program deeper into its portfolio of high-value rotating equipment. According to the company, with the monitoring system installed on 32 units, it was planning to incorporate more than 100 other assets into its Protean solution this year. EP