Ashland
            

Ashland Manufacturing is a $5 billion US-based provider of specialty chemical solutions. Employing 7,000 people worldwide, Ashland consists of two commercial units: performance materials and specialty (pharmaceutical) ingredients which their plant in Doel, Belgium handles.

Shifting towards higher added value and lower throughput

Ashland has been manufacturing construction materials since the seventies. Through the years, the plant has slowly changed its product focus towards personal care and pharmaceuticals. Naturally, transitioning its plant towards an entirely new market has sparked fresh challenges. For Ashland, it required a shift towards achieving higher added value, lower product throughput, and more control over its production processes. Tough compliance was also new territory—with increasingly strict GMP standards relating to pharmaceutical product quality.

Where people, processes and tooling come together

Ashland turned to TrendMiner’s self-service industrial analytics solution to get actionable insights out of its processing data. The platform appeared to be able to fulfill all of plant’s requirements and more: It could analyze automated plant production data in near real-time. It could visualize this data so that the team could apply proven methodologies such as Six Sigma and the DMAIC cycle. And most importantly: product engineers could use it with familiar tooling like computer aided engineering (CAE) and advanced analytics without needing a data scientist. “One of the important questions we asked ourselves” says Jan Meireleire, “was how we could transform large quantities of data into something that our product engineers could actually use. And—importantly—develop new ideas from.”

Ashland were running AspenTech IP.21 on a dedicated server to capture and store its historical data. TrendMiner’s self-service analytics platform was simply put on top of it and was deployed fast. Soon, engineers were able to analyze and monitor the data with powerful new insights. The following use cases show how Ashland solved previously “unsolvable” production issues, enhanced its quality control, and increased GMP production throughput.

Data analysis: stabilizing production by thinking outside the (data) box

One important goal in the transition of the Belgian plant was to stabilize production. Here, TrendMiner helped engineers to discover which factors in the process were influencing the quality of the finished product. “Interestingly,” says Jan, “these were often quite different from the ones we would have expected.” In this use case, TrendMiner identified the most stable production runs by filtering out data from specific products. While all the relevant parameters in these production runs turned out to be stable, a search of the entire database revealed significant factors right at the very beginning of the process. These so-called “influence factors” assisted with root cause analysis, helping engineers to determine the unknown correlations upstream. Thanks to TrendMiner’s easy to use dashboards, process engineers didn’t need to extract tons of data and figures into an Excel spreadsheet. Nor did they spend time manipulating data. Now these factors have been eliminated for a more stable production process.

Data monitoring: running a smooth plant

TrendMiner also comes with monitoring capabilities that enable process and production engineers to raise certain ‘red flags’ and prevent incidents from happening in the future. Based on patterns, engineers can highlight “Golden Fingerprints” and are notified when an incident occurs. In one instance, the process engineering group was notified of a specific event. Further analysis showed it had occurred several times in the past—quite a revelation! Not only were they able to solve the incident without lots of manual work, but they were also able to prevent similar ones from ever occurring again.

An efficient future

Ashland’s new self-service industrial analytics platform has boosted operational efficiency, compliance and quality control significantly across the plant. “Now,” explains Jan, “we are able to fix previously ‘unsolvable’ production issues and have seen an increase in GMP production from 70% to 95%.” No wonder Ashland plans to continue moving forward with self-service industrial analytics. This will mean focusing more on real-time monitoring and real-time prediction of asset performance going forward. Meireleire comments: “We wouldn’t be where we are today in over-achieving our goals without TrendMiner. This level of analytics on today’s process data will take our plant into the future.”