Rufen Sie uns an: +49 6221 427894-0
Industry: | Pharma, mid-sized |
Region: | DACH, Europe |
Process: | Produces active pharmaceutical ingredients (APIs) – manufactures various products in batches |
Challenge
Increase production capacity and debottleneck batch process in a reactor.
Solution
Benefit
The customer needed more production capacity, and were aware that capacity was available within the batches and steps - but they did not know where. They also wanted to know why there was so much spread in production time for different batches.
However, process engineers who understand the process are not data scientists, and vice versa. There can be concerns with using data scientists without sufficient information upfront and because of the investment required. Therefore, an easy-to-use tool – better than Excel - that allows process engineers to perform a pre-analysis would be extremely useful.
A Bilfinger project manager who understood the process but is not a data scientist was given a year’s worth of time-series data for 10 sensors. He easily used point and click to assemble and display a stack of trend graphs, for the four sensors that interested him.
By combining his knowledge of the process with his visual inspection of these graphs, he could make assumptions about the process steps; he confirmed with the customer that these assumptions were correct. He then handed the data to a data scientist, who developed an algorithm that identified that they had about 45 per cent spare capacity. This approach can also be used by the customer’s process engineer.
BCAP (Bilfinger Connected Asset Performance Platform) integrates data as a single source to unlock hidden potential. TrendMiner, as one of the benefits of using (licensing) BCAP, facilitated a very easy, low-cost route to providing a first root cause analysis. It allows process engineers to search past process behaviour, and achieve a diagnosis by instantly finding similarities and patterns.