According to Miguel Sison, senior reliability engineer for Chevron Corp. USA, maintenance and reliability can join other fields that have "really benefited" from digital transformation by extracting insights and driving decisions from robust data analytics programs using business intelligence (BI) software. It's just a matter of using the same data and the right data tools.
One such tool is the Weibull tool for predicting the probabilities of pump failure, a reliability analytics tool that is web-based and easily shareable, Sison said.
"The potential user will click the link on the email to get permission to access it, and then they can go ahead and use it. Work order and asset performance management data are updated automatically," he said. "The report extracts data from the Oracle server. In this case, it goes directly to the [power drive] services and populates the report."
It is necessary to use "a bit of coding" for some reports, Sison said, speaking as a member of a panel at the recent 2020 AFPM Summit.
"We use the R language, and for this one, it's the only coding skill involved," he said, adding that the R language is relatively easy to learn.
Sison said he believes this tool takes reliability to a new level.
"Before, we were constrained by the builtin capabilities of our computerized maintenance management system (CMMS). We can use all that stored data, whether it is in our CMMS and all those hundreds of databases where we store data," he said. "We can come up with inside stuff that otherwise would have been hidden for years in our existing CMMS."
Sison explained that it is possible in most cases to bypass doing the manual analysis in Excel.
"We just get the data from the server, feed it to our report and do the analytics from there," he said. "The reports that result are static, but people can use it as an actual tool that is interactive and flexible so they can configure it to their own needs."
Sison emphasized that reliability was previously highly dependent on IT support staff to do the reports or to develop the tools that were needed.
Now, with current BI tools like Weibull and Power BI, domain knowledge can be combined. "Folks who are already knowledgeable in that area and have them create the reports themselves," Sison said. "So we have the merging of domain knowledge and data analytics capabilities."
Weighing the benefits
Dan Oliveira, process safety engineer for Flint Hills Resources, joined Sison on the panel and noted the benefits of implementing BI software to generate process hazard analyses (PHAs).
"In operations, we were able to locate credited safeguards that protect us from bad data," Oliveira said.
Oliveira explained this option enabled operations to view high-severity consequences inside a unit and develop awareness on responding to alarms.
"Some of those safeguards were relying on operations' response to alarms to protect us," Oliveira said. "Through that data, operators were able to focus on their unit and what the alarms were protecting them from."
The BI technology provided operators another reliability benefit regarding scheduled maintenance on safeguards and instrumentation.
"They know the specific release devices they are relying on, and they can make sure they are on their maintenance schedule," Oliveira said. "We hope that this can help them if they have to bypass a safeguard before doing the work. It enables them to analyze the risk and study what the safeguard is protecting us from, so we can assess the risk before somebody has to perform maintenance.
"The tool has also helped the process safety management team develop metrics on PHA data and analyze the consistency on our PHAs, because now we can select multiple PHAs and compare them next to each other."
Abbas Dhalla, manager of manufacturing at Chevron Corp. USA's Reliability Center, moderated the discussion.