Proven Results: Eli Lilly

RESULTS

• Reduced first loop engineering cost by magnitude of 10
• Reduced training time
• 25% reduced configuration time
• 50% reduced system installation and check out time
• Reduced implementation time
• Reduced validation time/cost

APPLICATION
Batch pharmaceutical bacterial fermentation

CUSTOMER
Eli Lilly and Company. In 1982, Lilly’s Humulin product (biosynthetic human insulin) became the first pharmaceutical product in the world made by bacteria genetically engineered through recombinant DNA technology. Other well-known Lilly pharmaceuticals include the antidepressant Prozac and the antibiotics Ceclor (cefaclor) and Vancocin (vancomycin hydrochloride). As a point of history, the company was the first to mass produce penicillin using fermentation technology.

CHALLENGE
Until recently, process automation at Lilly usually involved dedicated, single-product manufacturing facilities validated to meet FDA regulations and then run as long as possible without change. Large- and medium-size facilities might be run by distributed control systems (DCSs), and smaller areas by programmable logic controllers (PLC).
Pharmaceutical manufacturers are reluctant to change controls because to do so requires revalidation, a costly effort due to lost production and intensive documentation. Lilly’s medium-size Humulin fermentation production area, for example, has run for 15 years with the original controls. Although the production equipment has many years of life left, spare parts for the controls are becoming harder to find. The automation and some of the instrumentation must shortly be replaced and the controls revalidated. Systems from Fisher-Rosemount and Foxboro, as well as some systems designed in-house, have all been used at various times since the early 1970s.

SOLUTION
Flexible manufacturing a goal
Lilly’s approach to pharmaceutical manufacturing is moving away from such dedicated facilities. The pace of product development is accelerating, especially in efforts to target specific diseases with tailored products derived from biosynthetic reactions. The company is now building facilities with smaller, more flexible process areas that can be linked in various configurations or switched from one product to another. Process revalidation is still an issue, but the advantages of production flexibility outweigh the cost. Up-front design using S88 (ISA S88 batch control standard) modularity also helps reduce costs.
Scalable automation ideal
The move to install flexible new facilities, and also to upgrade numerous older facilities, is altering the way Lilly approaches automating its processes as well. 
Until recently, the cost of the first channel in an automation system was in the range of $500,000. Today, new scalable automation cuts this cost to $50,000 or less. A DCS can now be considered for small applications that just a year ago would have required hybrid PLC/operator interface (OI) equipment. Today’s scalable automation has been a proven performer for medium-size production facilities.
To test the suitability of scalable automation, Lilly temporarily disconnected a portion of a legacy DCS system running an advanced bioreactor inoculant tank in an R&D pilot facility. 
The fermentation research facility was selected for several reasons. First, we avoided putting product at risk as we might in production. Second, success in a pilot environment should translate to success in production because production operations are more stable. Third, success in a fermentation operation should make for success in chemical operations. Bacteria are unpredictable, making fermentation a bigger control challenge compared to chemically
building pharmaceuticals. Last, a lack of control flexibility becomes readily apparent in a fermentation pilot plant where processes change constantly, control modifications must be made, and instrumentation is more extensive.