Bio g defense in depth presentation
0 likes609 views
The document discusses the strategy of "Defense in Depth" for increasing the robustness of biomanufacturing systems. It involves introducing redundancy to mitigate the impact of failures and ensure the system can tolerate variability and failures. The Bioproduction Group advocates using their modeling and simulation software to quantitatively evaluate different sources of variability and analyze how increased redundancy through approaches like scheduling flexibility and additional equipment can improve robustness. A case study example from Genentech showed how robustness analysis allowed optimizing existing equipment usage over installing new equipment.
Bio g defense in depth presentation
- 1. © Bioproduction Group | www.bio-g.com 10/24/2013 Defense in Depth Increasing Redundancy in Biomanufacturing Facilities October 24, 2013 Rick Johnston, Ph.D. - Principal and Founder Gary Wright - Sr. Account Director gwright@bio-g.com
- 2. © Bioproduction Group | www.bio-g.com “Defense in Depth” •Strategy for increasing the robustness of manufacturing systems •Introduces redundancy in manufacturing to mitigate the impact of failures •Key idea: system must be tolerant to failures •Implemented in many manufacturing environments and high risk operational settings (such as jet-liners) •When used successfully, avoids disruptions while maximizing likelihood of on-time batch release 10/24/2013 1
- 3. © Bioproduction Group | www.bio-g.com Bioproduction Group Founded in 2007 with an exclusive focus on Biomanufacturing Operations Primary goal of improving Quality, Productivity, Flexibility and Operations in Biomanufacturing World Class “real time” data collection, modeling, and simulation software Technology Assisted Knowledge Generation Tool Specifically designed for the unique needs of Analysts Improving Quality, Productivity, Flexibility and Operations at the World’s Largest Biomanufacturers
- 4. © Bioproduction Group | www.bio-g.com Why “Defense in Depth”? •Current biomanufacturing systems exhibit significant operational variability: –Batch-to-batch titers –Manufacturing times –Number of deviations per batch •There are also still a relatively high number of contamination events in the industry •The industry currently “hides” these issues behind large quantities of safety stock inventory and idled capacity 10/24/2013 3
- 5. © Bioproduction Group | www.bio-g.com Variability is seen throughout the manufacturing process. 10/24/2013 4 WFI Demand During Rituxan 3.5 rpw 2/09/09 to 3/(PI Data: tank level drop plus distillate flow sums) 0 200 400 600 800 1000 1200 1400 0 24 48 72 96 120 145 169 193 217 241 265 289 313 337 361 385 409 434 458 482 506 530 554 578 602 626 650 674 698 723 LPM WFI consumption (L/min) Cadence of batches in production fermentor Even run-rate cadence implies a ‘cyclic’ pattern of usage But, pattern of consumption is NOT cyclic
- 6. © Bioproduction Group | www.bio-g.com Temperature bands in a freezer 10/24/2013 5 Batches lost due to temperature malfunction
- 7. © Bioproduction Group | www.bio-g.com Lactate levels in bioreactor: significant batch to batch variability 10/24/2013 6
- 8. © Bioproduction Group | www.bio-g.com Non-stationary processes (i.e. process drift) 10/24/2013 Most modern biomanufacturing data exhibits both significant variability and significant process drift. CIP Times, 2002 - 2008 Hours
- 9. © Bioproduction Group | www.bio-g.com Comparative variability: semi-conductor vs. biopharmaceuticals Semiconductor Biopharmaceuticals Variation in per batch output (lower is better) * 3% ** 30% Number of deviations per batch 2 80 10/24/2013 8 * Standard deviation of performance per chip, http://spectrum.ieee.org/semiconductors/design/the-threat-of- semiconductor-variability ** Standard deviation of bulk manufacturing quantity per batch, internal Bio-G data, based on Mab production. The data suggests that biopharmaceutical manufacturing exhibits more significant process variability than other industries
- 10. Designing “Defense in Depth” into biomanufacturing systems
- 11. © Bioproduction Group | www.bio-g.com Traditional Risk Mitigation vs. Defense in Depth Traditional Defense in Depth Core Philosophy “Fixing each issue in turn will enhance reliability” “Avoid inevitable issues from affecting reliability” Focus Specific problem or issue identified Holistic view of manufacturing process Data Batch report or historian data for the issue Entire manufacturing system Method Root-cause identification and remediation Process as “Black Box” Tools Ishikawa, SPC, regression modeling, correlations, manual effort Discrete Event Simulation, automated analysis 10/24/2013 10 These two approaches are highly complementary
- 12. © Bioproduction Group | www.bio-g.com “Defense in Depth”: elements of successful risk mitigation 10/24/2013 11 Process Robustness Scheduling Redundancy “Surge” Capacity Inventory Buffering Equipment Redundancy
- 13. © Bioproduction Group | www.bio-g.com “Defense in Depth”: How •Bio-G has performed defense in depth analysis for more than 7 years for leading biopharmaceutical manufacturers •Our approach focuses on a data-driven approach to increasing robustness (rather than a consensus driven approach, relying on expert opinion) •Process Robustness: identify variability in processes and improve the systems that manage that variability. Includes items like unplanned maintenance and process restarts. •Scheduling Redundancy: introduce ‘holes’ in the schedule where significant variability occurs and that variability has a large impact •“Surge” Capacity: create the ability for parts of the facility to be able to ‘catch up’ when delayed or recover quickly due to an outage •Inventory Buffering: place intermediate WIP or raw materials such that they allow optimal recovery due to a failure 10/24/2013 12
- 14. Demonstration: How we model process robustness •Using data from automation systems / historians •Creating a model of the facility •Creating an automated robustness analysis
- 15. © Bioproduction Group | www.bio-g.com “Take home” messages 10/24/2013 14 •Increasing redundancy requires us to probe a manufacturing system, ‘imagining’ the effect of different kinds of failures •Most of the time, failures will have little or no impact to the metrics we care about (throughput, overtime etc.) – but the 5% of those that do matter are critical •An automated evaluation tool can be used to evaluate the impact of these failures •Single variable analysis shows us some impacts, but the best kinds of analysis look at multiple factors at the same time (DOE approach) •We must also look at multiple replications (i.e. repeating the same experiment multiple times) to ensure the answers are consistent
- 16. Case Study •From: Expanding Production at Biologics facilities: Effective strategies and Planning Ken Hamilton, Genentech Oceanside, Biomanufacturing Conference, Boston, June 27-28 2013
- 18. © Bioproduction Group | www.bio-g.com Installing an additional CIP skid produced the same result as optimizing existing equipment 10/24/2013 17 Confidence Histogram - By Resource 0 5 10 15 20 25 30 3.355 3.379 3.402 3.426 3.450 3.478 3.503 3.528 3.554 3.580 Run Rate (rpw) Number of observations 3 Upstream CIP Skids Upstream CIP Cycle Reductions No difference in the run rate and distribution of probable run rates between case for additional skid vs optimizing exiting skid
- 19. © Bioproduction Group | www.bio-g.com “ROBUSTNESS” ANALYSIS •Examines the effect on run-rate of delays in manufacturing operations •Goal: to allow for a robust schedule that, despite inevitable delays, will still allow us to reach our production targets •Robustness analyses look at varying levels of delays, typically from 1-8 hours (8 hours being an entire shift) •Can also be used to analyze the ‘white space’ available for preventative maintenance and calibration activities •Gives engineering groups targets for further improvement and areas to enhance operational efficiencies Credit: Expanding Production at Biologics facilities: Effective strategies and Planning Ken Hamilton, Genentech Oceanside, Biomanufacturing Conference, Boston, June 27-28 2013
- 20. © Bioproduction Group | www.bio-g.com Credit: Expanding Production at Biologics facilities: Effective strategies and Planning Ken Hamilton, Genentech Oceanside, Biomanufacturing Conference, Boston, June 27-28 2013 * No manufacturing specific data shown. Graphs show sample data only.
- 21. © Bioproduction Group | www.bio-g.com RISKS AND “WATCH OUTS” •Finding the optimum of likelihood of attaining target sustained capacity increase, cost and any shutdown durations is key –Find balance between optimizing existing equipment versus installing back up systems •Ensure capacity increase projects are always linked back to business needs –Business needs could change thru life of the project •Ensure scope of changes is thoroughly defined at the outset •Need to ensure operations groups and teams remain fully engaged thru life of project –Ideally transition project to an operations group toward end of implementation phase •Develop accurate cost estimates early in the project –Avoids recycle Credit: Expanding Production at Biologics facilities: Effective strategies and Planning Ken Hamilton, Genentech Oceanside, Biomanufacturing Conference, Boston, June 27-28 2013
- 23. © Bioproduction Group | www.bio-g.com Key idea: do robustness analysis in real-time and use that to schedule 10/24/2013 22 Real-time analysis and optimization is critical to achieving ‘Best-in-Class’ performance
- 24. © Bioproduction Group | www.bio-g.com Collect feedback on robustness from Outlook (automated toolset) 10/24/2013 23 Allow manufacturing to instantly react
- 25. © Bioproduction Group | www.bio-g.com … or using smart phones that are on the manufacturing floor 10/24/2013 24 * http://www.cocoanetics.com/2011/12/myth-busted-iphones-wont-work-with-gloves/
- 26. © Bioproduction Group | www.bio-g.com These toolsets have a real impact on manufacturing 10/24/2013 25 Reactionary Expediting 86 13 Year 1 Year 2 8 86% fewer 91% fewer Before Hours Spent (Mfg + Scheduling) 1837 1322 Year 1 Year 2 900 28% less 51% fewer Before Adherence to Plan 76% 88% Year 1 Year 2 98% Before
- 27. © Bioproduction Group | www.bio-g.com Conclusions •“Defense in Depth” is a holistic approach to designing redundancy into manufacturing systems •The approach requires toolsets that ‘understand’ variability and the impact it could have •Rather than ask for people’s view on risks, it uses a data driven approach that is based on past performance •Designing robust systems is a complementary approach to root cause analysis •These approaches do not require massive investment in infrastructure •When used, it can have significant benefits to the business for metrics like reactionary expediting and performance against plan 10/24/2013 26