Eli Lilly was faced with the challenge of finding additional capacity at one of its manufacturing facilities to accommodate a new process. Through a partnership between its process modeling and manufacturing groups, the company identified combinations of process adjustments and cycle time reductions that could provide the required capacity with no additional capital spending.
The modeling environment and toolset maintained by Lilly’s process modeling group, part of its central engineering group, enabled the rapid development of an overall process model and evaluation of over 150 unique combinations of process alternatives, ranging from cycle time improvements to addition of capital equipment.
The results were delivered within the same modeling environment the manufacturing group already uses for emissions monitoring. This effort has demonstrated the superior capability of the modeling environment compared to the spreadsheet-based systems now employed for other tasks — and has renewed interest in using it more widely for process improvement and technology transfer.
The facility has a number of batch unit operations that several processes can share. Lilly wanted to use the equipment for a new process. The challenge was to reduce the campaign of an existing process for manufacturing an active pharmaceutical ingredient (API) from approximately 40 weeks to 26 weeks or less, allowing the new process adequate time on the equipment. The project also had a tight schedule because the new process was to be moved into the plant within a year. If the existing process couldn’t be modified in time, it would have to be relocated to larger equipment in another building on the site or even to another plant — both options requiring considerable time and expense.
Meanwhile, the manufacturing engineering team was pursuing several process improvements, including optimizing current unit operations, telescoping two process steps into one, and raising process yield.
The project team at the manufacturing site recognized the complexity of the effort and collaborated with the process modeling group to explore whether modeling tools could assist with choosing the optimal production options.
A high-level simplified process model was developed to capture the overall cycle time and yield of the relevant process steps. This model didn’t include all materials specified in the recipe. For example, water was used in place of specific process constituents; this allowed for both rapid model construction and acceptable accuracy (because the model isn’t used to carry out an overall material balance). However, the model structure would easily incorporate materials and other particulars to conduct a future detailed analysis if so desired.
Individual operations such as vessel charges, heating/cooling, material transfers and separations weren’t modeled. Instead, higher-level “unit procedures” (e.g., extraction or crystallization) were. Each unit procedure typically consists of just one or two operations, allowing modeling of the overall cycle time and yield for that procedure. This approach enabled rapid model development and, by using the integrated production planning and scheduling capabilities, quicker simulations of campaigns.