Emerson Exchange 365

In early 2013, Novartis Vaccines began planning a major overhaul of the process automation system at one of its bulk manufacturing facilities in Italy. The goal was to adopt a fully integrated process strategy, one that would improve flow consistency by moving from sequential unit operations to single production lots. The company had high hopes for the three-year project, as Gabriele Ricci and Massimo De Francisci of Novartis explained this week at the 2013 Emerson Global Users Exchange in Stuttgart.

“We wanted to simplify quality control by combining production data according to a unique batch and lot ID instead of generating multiple batch reports for each execution phase,” Ricci said. “We also sought to adopt collaborative operations, which we planned to do by integrating real time process data with raw transactional and business data. We expected that if we increased ownership in the control system, we could drive future changes and expansion as part of a consistent evolutionary strategy.”

“We based the modernization project on two central pillars,” De Francisci added. “First, bringing together an integrated team of stakeholders in engineering, IT, manufacturing and quality control; and second, implementing a detailed front end engineering design (FEED) phase. By doing this, we hoped to identify key performance indicators and mitigate potential risks as early as possible.”

The six-month FEED process followed a tightly structured methodology: The first step involved collecting appropriate information and data through onsite stakeholder interviews. A detailed project execution strategy was then formulated in step two. Step three focused on refining the business processes that would be integrated with the new automation system once it was in place. Step four involved defining key requirements for the new system. The preliminary roll out plan was drawn up in step five; step six was reserved for addressing organizational change issues; and governance documents were drafted in step seven.

“In addition to the usual FEED checklist, the project leads came up with a number of 'extended deliverables' to help achieve our long term targets,” said Ricci. “The first of these was standardization. We wanted to adopt common best practices across the company to reduce testing and maintenance time and improve validation procedures. Next was an application software prototype with field simulators able to identify and correct any potential issues with execution or the human-machine interface. Then there was operator and engineer training, which allowed our personnel to gain confidence with the new system and provided an opportunity to create a baseline of process configurations for future application development. Finally, we made it a priority to re-use existing field connections to reduce implementation time, prevent human errors, and streamline testing procedures.”

The project team pushed the FEED evolution even further, however. “Because of the complexity and duration of the project, we saw the need to envision how the system will be used three years from now,” De Francisci explained. “We asked ourselves whether the technology really assists the operator in such an automated environment, or if it's the other way around. This is how we arrived at the idea of the operator as the fulcrum of the process. The key is to improve operator awareness with a fully human-centric design approach, using the DeltaVTM process automation system as the 'glass door' on the production process. To do this, we need to move from transactional data management toward more of an emphasis on usability, mobility, and real time process control.”

Going above and beyond normal expectations during the FEED phase has paid off for Novartis in many ways, both organizationally and fiscally. These include faster implementation of technology; a more agile development process; building stronger stakeholder relationships; and ensuring that the design meets the project timetable, budget, ROI and operating cost criteria—before diving into a big project. Rethinking FEED has also reduced human errors and cut down on document review and data entry time. It has also increased system robustness, reduced demand for spare parts, and improved the overall reliability of the process.

“This novel approach to system design has allowed us to make use of business intelligence in ways never before possible,” Ricci concluded. “While the current generation of manufacturing platforms was designed for functionality, the next generation must be designed for analytics as well. We should approach industrial automation as part of a much broader vision in the future.”