Smart Sensors and Data Management Solutions for Modern Facilities

13 Min Read

AdobeStock_131421720-300x200.jpgBioprocess manufacturers continue to seek technologies for increasing productivity and shortening timelines from discovery to commercialization. Innovations such as high-throughput systems, automated platforms, and the latest clarification systems all have made processes efficient and robust. And with the increasing adoption of quality by design (QbD) principles, including the use of process analytical technologies (PAT), biomanufacturers are mitigating the risks of errors in their operations better than ever before. A critical part of mitigating risk is gathering meaningful process data and then relating those data points to information about product quality. Smart facility management starts with the implementation of solutions that can relate those data to decision makers quickly — ideally in real time — as part of one overall network.

To gain perspective on modern technologies and design approaches in bioprocess facilities, I spoke with Matt Roesch, senior director of life sciences at JLL Life Sciences.

Sensors for GMP and Non-GMP Areas
Biomanufacturers are seeking ways to reduce time to market. What innovations are likely to help them achieve that goal? Having spent over 30 years in the life sciences industry, I always have approached this question by considering the impact that some technologies have had on both good manufacturing practice (GMP) facilities and non-GMP facilities. Overall, the innovations that have been the most active in the industry are smart sensors, which are leading to uses of predictive analytics.

Sensors_photo-1024x280.jpg

Examples of current sensor technologies: (left to right) Sonotec’s Sonoflow CO.55 noninvasive flow sensor, Broadley James’s SingleSense SU800 series pH sensor, and PendoTECH’s single-use port plate pressure sensor for flexible containers

Many companies and consumers are facing different options among newer technologies and are seeking applications for them in industrial sectors, including life sciences. A common example is of new sensors that measure temperature and humidity. They have applications both for GMP and non-GMP settings in modern facilities, but they also can be great for troubleshooting in existing facilities with Wi-Fi capabilities. Extending Wi-Fi capabilities to remote facilities is less expensive than having a centralized building or utility management system. Wi-Fi also can retrieve good data for troubleshooting or extending capabilities, and that capability has opened up the playing field for a lot better data at lower costs.

For a GMP environment, few of the newest technologies are validated as required. So you need be careful when thinking about using new sensor technology in a GMP area. For example, suppose you want to know temperatures in a cold room or other environment. As soon as you set up a system for that, the data you gather could have a GMP quality impact. So you need to think about what you’re going to do with those data, especially if you obtain a reading that falls outside of your validated parameters.
Many environmental temperature sensors can gather real-time data. I have had some experience with sensors that generate discrete data about every 10 minutes, but the newest sensors now are gathering data every five seconds. Some pH sensors take readings over short periods, but those data may or may not be meaningful, depending in part on whether you are obtaining batch results or real-time results.

Data Management
How do the data generated by new sensors affect how operators do their jobs? Once you start gathering data, and you’re validated for certain parameters of pH, for example, you need to decide what to do if the data point falls outside of those parameters. Is that result a problem? Do you have to raise an incident and then investigate a deviation? Do you have to have product put on hold? You need to go back to look at the science to determine what that supports. Innocent acts in a facility can lead to significant outcomes if you aren’t thoughtful up front from a GMP standpoint.

In terms of facility design, compare the GMP space of your pilot plant and manufacturing space with your office and even your research and development space, for example. The environmental needs are different for all of those spaces, and the division is a fundamental distinction. I’ve seen some people innocently get themselves in a potential trap by not taking those factors of GMP and non-GMP spaces into consideration.

From the BPI Archives

Ratcliff A, Preisig C. Advances in Sensor Technology Improve Biopharmaceutical Development. April 2013.

Thomas A, Munk M. Meeting the Demand for a New Generation of Flexible and Agile Manufacturing Facilities: An Engineering Challenge. December 2015.

Harrison R. Strict, But Flexible, Industrial Automation for Biopharmaceutical Manufacturing. January 2016.

Whitford W. The Era of Digital Biomanufacturing. March 2017.

Godfrey JT. Extend the Life of Your Facility: Flexibility Allows for Biopharmaceutical Process Innovation. January 2018.

Beaudoin H, Meirzon S. Partnerships for Progress: Supplier Perspectives on Facilities of the Future. April 2018.

Center A. Facilities of the Future: Intelligent Design and Control Enable Quality, Efficiency, and Good Citizenship. April 2018.

Montgomery SA. Flexibility, Automation, and Leadership: Drug-Sponsor Perspectives on Modern Biomanufacturing Facility Design. April 2018.

Kopec D. Using Modern In Situ Analytics and PAT for Automated Feedback Control of Critical Process Parameters. November 2018 ebook.

Beri BG. Opportunities for Modern Robotics in Biologics Manufacturing. April 2019, featured report.

Any time you gather data in a validated space, you’re potentially making those data part of the official record. Some new sensors are being used for diagnostics, and some are available at relatively inexpensive costs for use in commercial buildings. But their software and hardware are not validated to the Food and Drug Administration (FDA) standards. So even though the data are real, they are not being gathered through validated sources.

You can get into difficult situations when you gather data in validated and nonvalidated spaces without factoring in the state of that area. For example, when you’re conducting a test outside of a validated state, for which the data gathered will be used solely for making adjustments, then you could perform a validation separately. That might be okay. In such a scenario, it comes down to the quality perspective from each organization. But when you’re doing that same task in a validated area, then you have to be aware that the data being taken from sensors could be used for records. Just because we can gather data inexpensively and collect more than we ever have before doesn’t mean that we should.

Some of us who have been in the industry a long time have had to face these questions: What are we going to do with all the data gathered? What kinds of decisions are we going to make? You can experience data overload with the capabilities of new sensor technologies available right now. Many engineers can get themselves in the same trap of wanting more and more data for different parameters, whether those data are meaningful or not.

Some central systems can be set up to gather information. But then when you augment that with a lot of Internet of things (IoT) technology, then you need to determine how to analyze and filter those data, which is the case for many large and even medium-sized companies. They often invest significant money with major manufacturers of utility or building management systems. Those systems depend on particular applications, which may or may not be validated. If a system is validated, then some biomanufacturers want to have everything tied into that system. If those applications are not validated, then companies have data historians for analytics. Even for general facility parameters, you need to determine how to integrate new technologies into your existing technology platforms, hardware, software, data, and data-storage systems.

Understandably, biomanufacturing companies are reluctant to seek out different technologies unless there is a specific application that their current platforms do not meet. If you’re trying to bring in additional new technology, how do you integrate that into your existing technology platforms? One solution that my company has seen with some clients is using predictive analytics and vibration data. With traditional vibration data from sensors, data points can be a fairly extensive part of a mainframe central system. By contrast, some startups and newer companies gather data at a fraction of the cost and at capacities not available with a mainframe large system. Generally, if you’re gathering that type of data, you are doing so to support valuable pieces of equipment.

Those data also may be critical to sustaining environmental operations such as maintenance, although such operations are not likely to be in critical GMP spaces. We’ve seen some traction in this area from clients who are feeling that there’s real value and insight in gathering data from such operations because doing so provides real information about their operations.

Advancing Sensor Capabilities

Biomanufacturers are seeking state-of-the-art sensors for use in advanced designs such as closed, automated, and integrated process schemes. Increasingly, companies are seeking sensor devices that are noninvasive, flexible, and highly sensitive and that can be used in single-use and multiuse applications. Some inline sensors now available can transmit real-time data to a wireless device (e.g., laptop or tablet) through Bluetooth technology.

The latest technological innovations are part of the push toward digital manufacturing and what has been referred to as “Industry 4.0.” The Internet of Things (IoT) and the Industrial Internet of Things (IIoT) are novel technical paradigms that allow instruments, complex information systems, and process equipment (“things”) to be connected and communicate through the Internet. Connecting IoT sensors to smart devices can obtain data from pumps, reactors, laboratory equipment, and other machines and devices. Those sensors also can be connected directly or indirectly to IoT networks. The ideal set up would be to establish a wireless sensor network (WSN) within an IoT structure.

What are some common needs you’re seeing from your biomanufacturer clients? Clients are seeking to support better decision making at a more reasonable value than they have historically. So many have large systems that are robust, but the cost to add more capability and even to keep what they already have is a significant annuity or requires significant stepwise investments. They’re asking us to help them advance their thinking and decision making at a much better value then what they had before instead of just gathering tons of data.

So this involves not only technologies, but certain strategies for approaching those data? Absolutely. Biomanufacturers need better data analysis to make informed decisions early and quickly in their processes. They want us to help guide them toward the right technology for their specific environments and goals so that they can return to their primary responsibilities.

Biomanufacturers also are needing to handle big data and correlate that with their operations and product quality. Do you agree that it’s a general problem? Yes. It’s similar to modern technologies we have now that we thought would make our lives much more productive but have added much more work volume. We’re right on the cusp of that or in the middle of it in some cases in this particular area right now. And I would just caution everyone that there are great technologies out there, so you just have to take all factors into consideration and begin with the end in mind.

Wi-Fi Networks and Other Technologies
Are new technologies such as digital twins, model predictive control systems, and so on being implemented yet? In my experience, some of those technologies are more on the leading edge and will take a while to filter into existing biomanufacturing facilities. They would be more useful when incorporated into new facilities or when a company makes changes to existing operation or set of parameters.

Many large companies also have installed secure Wi-Fi or similar networks to prepare for future wireless technologies. Some biomanufacturers install guest networks (typically Wi-Fi, although not always) that allow outside connections for guests. Many new sensors are not set up to meet (or not capable of meeting) information technology (IT) security protocols, and that setup can be cost prohibitive. Some companies allow those types of technologies to connect on a guest network outside the internal, more secure network.

If you are modifying a building or a process, you have to consider how large a data “pipe” that you need throughout a building, but also whether you can host external devices and future systems and whether doing so could become problematic. For areas with industrial equipment and the like, setting up guest networks and Wi-Fi hotspots may not work well, so systems different from those used outside of those areas would be needed. Some companies use a hub plugged into their networks to handle transmissions in such spaces and to gather relevant process and environmental data.

For companies wanting to make processes smarter in existing facilities, what technologies are available in addition to sensors? The biggest change has been the installation of robust networks that host or allow external connections. For example, Bluetooth “beacons” are now considerations for biomanufacturing facilities. They are signal repeaters located throughout a building that can locate people and identify where they are through the use of a phone app. That application requires a robust infrastructure of Wi-Fi capabilities, but it is useful for locating badged and nonbadged people. Such companies have a platform through which personnel can connect, bringing value and connectivity to people throughout a space.

Predictive analytics technologies are another useful capability. These systems take data and provide a way for biomanufacturers to make decisions proactively instead of using data to search for the root cause of a problem after the fact. The direction at most companies has been to implement better coverage and secure better data with a smart analysis up front before failures occur.

Everyone is trying to determine the next major technology or approach that will be adopted widely by biomanufacturers. Technology has changed more in the past few years than it has in a long time. Going back about 20 years, we had central building systems — building management and utility management systems — and some of those are still in place at some facilities. But much change has taken place in the biomanufacturing industry. Technologies now on the edge of our capabilities are expected to accelerate speed to market further in the next three to five years. But there is still so much infrastructure related to these new technologies, facilities, and processes that it will take time to make those transformations and secure some degree of investment.

Corresponding author Maribel Rios is managing editor at BioProcess International; [email protected]. Matt Roesch is senior director, life sciences, at JLL Life Sciences.

You May Also Like