1-3-will-olp-honeywell-pmt-main
1-3-will-olp-honeywell-pmt-main
1-3-will-olp-honeywell-pmt-main
1-3-will-olp-honeywell-pmt-main
1-3-will-olp-honeywell-pmt-main

Cloud analytics unlock asset performance

June 10, 2019
PMT Director Will Olp explained how his facility realized 2% productivity gains in just six months.

“We were able to take the process information and apply it to the algorithms and data that were being streamed into the cloud.” PMT Director Will Olp explained how his facility realized 2% productivity gains in just six months at Honeywell Users Group Americas in Dallas. 

By leveraging cloud-based analytic technologies from Honeywell Connected Plant (HCP), a six-month internal initiative boosted productivity of machinery and processes running within Honeywell’s own Performance Materials and Technologies (PMT) division. The pilot program increased revenue and capacity by more than 2% at the PMT plant in Orange, Texas, demonstrating the value that today’s cloud analytics can provide toward recognizing hidden asset performance degradation and how assets interact with the processes they serve.

“We make low-density polyethylene,” said Will Olp, manufacturing director of the additives-and-chemicals group within PMT, in his presentation at Honeywell Users Group Americas 2019 in Dallas. “We were having more unplanned downtime than business could support. I had to have a solution, and I had to come up with a way to get this done quickly with the existing resources I had on-site.”

The plant had experienced several unplanned shutdowns over 24 months, an estimated $20 million negative impact, explained Olp. Performance analysis and reporting was lagging, manually gathered biweekly. Maintenance was highly reactive, repeatedly finding shutdowns were avoidable. And communication paths between engineering and reliability/maintenance were inefficient.

“The plant incurred several costly unplanned shutdowns,” explained Olp. “There were myriad operational challenges and worn plant infrastructure. We didn’t want to wait for the engineers to do the analysis. We partnered with HCP, and they have engineers who know a lot of the equipment. We were able to take the process information and apply it to the algorithms and data that were being streamed into the cloud. I started this process in May 2018, and we had a solution in place in September. In the last quarter of 2018, we received results.”

The project team combined reliability engineers, process engineers and operators. “Within Asset Sentinel, they’re now putting in what the root cause was,” explained Olp, “in addition to creating the algorithm from the data in the historian. We have 28 assets configured and 58 measured KPI sentinels.”

Uptime contributions

The process and asset monitoring data streams are unified into a single data-processing scheme. At the same time a duplicate stream is being sent to the cloud. “It takes a copy of the data historian and upfeeds it to the cloud,” explained Olp. Monitored equipment included reciprocating compressors, heat exchangers, boilers, reactors, off-gas spray columns, binary columns and deodorizers. Four areas of contribution came from the heat exchangers, deodorizers, boilers and an unidentified trade-secret asset.

“Heat exchangers will foul over time and need to be cleaned periodically,” said Olp. “The analytics looked at the trend of the fouling and when the next planned maintenance was, allowing for increased throughput.” This accounted for 16% of the contribution.

On the deodorizer, the new model alerted significantly earlier, and engineers used steam to clear fouling. “Had alerts occurred later, the column would have operated past the point where extra steam would be effective, thus shortening packing life and requiring shutdown and premature packing change-out,” said Olp. This resulted in 13% of the contribution.

Historical patterns revealed correlations and a prolonged maintenance interval in the unidentified asset. “By monitoring the seal, we were able to predict that a bearing would have trouble, and the operators were able to make adjustments to extend the life of the bearing to the next shutdown,” explained Olp. “Unique shift patterns were found through historical analysis of bearing vibration data, and shifts revealed pending failure of the nearby packing gland (seal), rather than the bearing itself. Previously the glands were a preventive maintenance item and replaced early.” This was the largest overall contributor, accounting for 55% of the gain.

Boilers improved in overall efficiency and operated within environmental regulations by increasing efficiency through better utilization of waste heat, resulting in better throughput, a 16% contribution.

While a 2.2% increase in revenue and 2.2% additional capacity were realized in six months, financial projections indicate breakeven in just 1.1 years annualized.