Batch isn’t just for breakfast anymore. Traditional process industry applications are evolving rapidly, yielding to new and improved ways of gathering more and better data to help optimize high-technology processes.
“InFusion was the only solution able to immediately bring intelligent information to the plant floor, and then incorporate that raw goldmine of data into a plant historian.” Bayer’s Noël Jans on the pharmaceutical manufacturer’s push to aggregate and convert its reams of under-utilized data into actionable intelligence.
In two industry-breakout sessions this week at the 2007 Foxboro User Group Conference, a “dynamic duo” from Bayer Antwerp showed how to extract useful data from process settings, while a scientist from the University of Massachusetts’ Bio Manufacturing Center (BMC) described how his facility is optimizing its new pilot batch reactor. Both applications employ Invensys’ InFusion enterprise control system to accomplish their goals.
Bert Baeck and Noël Jans work in Bayer’s Central Automation Group, which is responsible for 22 Foxboro I/A Series systems at nine Bayer plants. To begin improving their applications, the two engineers say organizers formed an advisory group to find and demonstrate superior technologies to Bayer’s production managers, and then show how these improvements could fit into existing applications.
“We had to learn to defend a large investment in, for example, a new alarm system, and show it was better than a customized system, how it would address maintenance, generate return on investment (ROI), and contribute to the application’s lifecycle,” said Jans. “To do this successfully, you need to make plant managers aware of the goldmine of information available in their application. We’ve found that only about 5% of the total process data generated is presently turned into useful, actionable information, and 50% of IT projects eventually fail because expectations weren’t thoroughly defined ahead of time.”
With an assist from InFusion, Jans added, the company’s Central Automation Group already has implemented several “low-hanging fruit” improvements at Bayer’s plants. However, lengthy discussions and incremental implementations also are needed to successfully achieve these gains. “We try to phase in improvements in a series of smaller projects, often three or four weeks long each, and then add up their results and successes after a few months, so we can keep the overall improvement project running. It also helps to have a ‘horizontal specialist,’ who can write up the value-adds in a project proposal across all the domains that we’re interested in--including control. Engineers today also need to add economics, math, and other technical skills.”
Baeck added that Bayer’s approach begins with a huge data collection effort, and then complex algorithms are applied to what’s been gathered. Success story details are compiled next, which may include throughput improvements or reduced operating costs. These short-term improvement projects can explore available data, and help users perform key performance indicator (KPI) spotting. This can help identify useful information on costs and workflow, as well as success factors that may not have been visible before. This means more accurate return on investment (ROI) evaluations for particular improvements, and better arguments for investing in them.
“InFusion was the only solution able to immediately bring intelligent information to the plant floor, and then incorporate that raw goldmine of data into a plant historian,” said Jans. “This not only enables ROI statement, but also breaks down boundaries between business and technology people and between and between IT and DCS engineers.”
Meanwhile, InFusion also is being installed in a pilot plant’s bioreactor at the University of Massachusetts Lowell, according to BMC’s Dr. Carl Lawton. The center received $50 million in initial state and other financing, and is expected to receive hundred of millions more as its missions and projects multiply as part of the Massachusetts Life Sciences Initiative.
“There hasn’t been a lot of process control in biopharmaceutical manufacturing yet, and so the joke is that potato chip makers are more advanced than us,” says Lawton. “However, there now are five interrelated capabilities to help biopharmaceutical companies transition from drug discovery to manufacturing. These include developing process development services, large-scale engineering facilities, education for company leaders and staff, applied research focused on critical bio-manufacturing issues, and networking existing industry and academic expertise.”
Lawton reported that one overall effort in his field is to get away from using the typical 1,000-gallon stainless-steel vessels and to focus on short-run, pilot production that can confirm proper process development and resolve commercial production issues. This allows developers to reduce costs of their initial engineering runs, and to conduct more of them. For example, Lawton said that BMC has been working with Nova Biomedical and its nutrient analyzer on a joint glucose feedback control system. Though also still in the pilot stage, this project could improve several biopharmaceutical manufacturing processes.
In addition, InFusion is being used to automate the IT functions on a 190-liter bioreactor that Wyeth recently donated to BMC. Control capabilities required on the reactor include an accurate and validate-able process control system, flexible and scalable functions that can be modified with minimal cost, compliance with FDA 21 CFR Part 11 and ISA 88, and supervisory control and data acquisition (SCADA) functions.
“We can use the system we’re building now with InFusion to make good laboratory practices (GLP) materials, and then move into good manufacturing practice (GMP) areas,” said Lawton.