Centre for Precision Engineering, Materials & Manufacturing Research,
Institute of Technology Sligo
A method to determine excessive energy cosuming batch manufacturing processes.
Batch manufacturing processes are extremely energy intensive. These processes can benefit from statistical analytical methods to identify production steps that are highly variable or contain outliers. Outliers, in this sense, represent singular batch processes that consume significantly greater energy than other observed identical batch processes. This work presents a case study on the analysis of manually recorded written batch manufacturing records from a pharmaceutical facility. The highly variable process steps and outliers are identified using boxplots. These are further analysed to identify causes, which include transcription error in data recording, parallel processing between manufacturing locations sharing utilities and heuristic based decisions made by plant process engineers. By identifying and acting upon these causes, the facility can achieve greater energy efficiencies and have a more sustainable approach to batch manufacturing.
GlaxoSmithKline (GSK), one of the world’s leading pharmaceutical health care companies, currently operate a high volume batch production process facility in Sligo, Ireland. In this facility, topical skincare products such as creams, lotions, gels and liquids are manufactured. Each production process requires a number of 3-phase asynchronous motors and different temperature range controls, which at times can exact significant demand on plant energy capacity to service multiple processes operating in parallel.
This work presents the analysis of manually recorded written batch manufacturing records. Therein lies the challenge posed to conventional statistical analysis. As batch records are currently recorded in writing, it poses a barrier to rapid identification of process variability and outliers. Outliers, in this sense, represent singular batch processes that consume significantly greater energy than other observed identical batch processes. As the number of batch manufacturing cases increases, so too does the complexity of the problem. The benefit in identifying processing steps that are highly variable is the introduction of standard operating procedures that reduce variability. Outlier identification allows for batch processes to be further investigated so that causes are identified and acted upon to prevent future occurrences.
Approximately 1000 batch manufacturing records, which were documented over a one year period, were digitised and transformed into a useable format for analysis. The duration of each processing step and the time between each processing step was calculated for every batch and product. The times were then plotted using boxplots. Taking this approach resulted in simple identification of highly variable process times and time between processes and outliers. Figure 1 presents the results for a single product “Z”. All product and batch information and process data have been anonymised. Process step “P1” represents the total time for each batch of product “Z”. The steps P2 to P8 represent each of the individual processing steps.
It is evident when viewing Figure 1 that highly variable batch processes and outliers are easily identifiable. Each letter identifies a batch of product “Z” that is considered an outlier. The outliers are of particular interest in these cases as they represent batch processes that consume significantly more energy than the other cases undergoing identical processes. It is also apparent which processing steps have more variable processing times than others. The transparent identification of outliers and more variable process step times allow for further investigation to identify causes. Typical explanations include transcription error in the written data records, parallel processing between manufacturing locations sharing utilities and heuristic based decisions made by plant process engineers. These can now be acted upon to improve energy efficiencies by introducing operating procedures. This will result in more sustainable manufacturing practices.
About Konrad Mulrennan:
Dr. Konrad Mulrennan is a postdoctoral researcher in the Centre for Precision Engineering, Materials & Manufacturing Research at Institute of Technology Sligo. Konrad’s PhD was partially funded by the EU FP7 “Research
for the Benefit of SMEs” programme on a project called Bio-PolyTec: Accelerating the Development of Bioresorbable Medical Devices. His PhD focused on the development of soft sensor technologies with application in extrusion processing of medical devices. He is currently collaborating with GSK for the Northwest Centre for Advanced Manufacturing (NW CAM) funded by Interreg VA. His role involves performing data analytics and model development to achieve energy efficiencies and more sustainable manufacturing practices. Konrad’s current research interests are in the areas of data science, predictive modelling and sustainable manufacturing.
IT Sligo established the Centre for Precision Engineering, Materials, and Manufacturing Research (The PEM Centre) in 2013 that is lead out by Dr. David Tormey. The Centre comprises of a team of multidisciplinary Principal Investigators, postdocs and a number of MEng and PhD students whose research outputs can be cumulatively summarized by the following: €10 million secured from competitive research funding, 250 peer reviewed publications, 2 patents granted, over 100 industrial projects completed including €1.6 million in funding directly contributed from industry, Graduated 11 MSc and 6 PhD students. The PEM Research Centre is actively engaged in a number of collaborative research projects with manufacturing industries including GSK, Abbott, Intel and Abbvie. The PEM Research Centre is also a designated Technology Gateway, funded by Enterprise Ireland, to support the applied research needs of the manufacturing industry, and are partners in the I-Form Advanced Manufacturing Research Centre funded by SFI and the Northwest Centre for Advanced Manufacturing (NW CAM) funded by Interreg VA.