Podcast: Play in new window | Download (Duration: 21:51 — 15.0MB) | Embed
Subscribe: Apple Podcasts | Spotify | Email | RSS | More
Process Analytical Technology, or PAT
, was introduced by the U.S. Food & Drug Administration back in 2004. The purpose of this guidance was:
…to describe a regulatory framework (Process Analytical Technology, (PAT) that will encourage the voluntary development and implementation of innovative pharmaceutical development, manufacturing, and quality assurance.
Raman spectrometers perform spectral analysis to enable consistent product quality, process intensification, and real-time control. The challenge with their use is to integrate their wealth of multivariate data into the control strategies to drive these performance improvements.
Boundless Automation liberates data to unleash the power of software to drive world-class performance. One example is Emerson’s solutions for PAT-based manufacturing, which helps streamline manufacturing processes by integrating multivariate data and real-time process models to manage critical quality attributes and predict final product quality during production. DeltaV Spectral PAT brings in real-time multivariate analysis to help enable fully automated manufacturing for improved speed-to-market.
In this podcast, Emerson’s Jorge Costa and Bruce Greenwald join me to discuss the application of PAT technologies not only within the Life Sciences but across many of the process and hybrid manufacturing industries.
Give the podcast a listen and visit the Process Analytical Technology and DeltaV Spectral PAT sections on Emerson.com for ways to help you drive innovation and improved performance.
Transcript
Jim: Hi everyone. I’m Jim Cahill with another Emerson Automation Experts Podcast. Process Analytical Technology, or PAT, is a known solution for in-line, real-time quality monitoring in manufacturing. Yet, one can debate to what extent it has been adopted across industries. Emerson has recently released DeltaV Spectral PAT to help the industry further progress in their PAT journey.
This podcast will revisit PAT’s value proposition, highlight the technology, and recommend a strategy to realize the full potential of PAT in manufacturing. Bruce Greenwald and Jorge Costa have been collaborating on PAT business development, namely in the Life Sciences industry. Bruce Greenwald is the DeltaV platform business development sales director based in Austin, Texas.
And Jorge Costa is a Life Sciences industry consultant based in Lisbon, Portugal. Welcome, Bruce and Jorge.
Jorge: Hi, Jim. Thanks for having me.
Bruce: Thanks, Jim.
Jim: Hey, it’s great to have you both. Why don’t we get started? Bruce, can you share a little bit of your background with our listeners?
Bruce: Sure. I’m a 1979 graduate of the University of Kansas where I majored in chemical engineering with an emphasis in process control and have spent the last 50-plus years in the automation space, working on originally our Provox platform. And then moving to the DeltaV platform back in the early 2000s.
Jim: Well, that sounds like a great backdrop and background there. Jorge, can you share your background as well?
Jorge: Sure. So I am, as you said, I’m based in Lisbon. So I studied here and did my career over here. I have a degree in chemical engineering in the nineties.
I joined Life Sciences industry in 99. I worked as an automation engineer for a few years, then went into automation management and corporate management. I spent a lot of time on process digital transformation and digital transmission strategies. And in the last two years, I joined Emerson as a Life Sciences consultant, as you said, and I’m also helping here about bringing the value message to our customers about the solutions from AspenTech and Emerson.
Jim: Well, that’s great. Thank you both for that background. Well, let’s get started and dive into the topic of PAT. Jorge, let me start with you, what is the value of PAT to the industry? And I guess what industries can it fit?
Jorge: Well, thank you for this question. So yeah, definitely, there is a big value, and there are different dimensions of this value.
I would say performance, compliance, throughput, and safety is what comes to my mind. So I can help break down and kind of give you some more context. So when we are talking about PAT, it’s about bringing the capabilities of the quality control lab into the production shop floor. And, you know, as a result of a process, of data modeling that starts in development space that can correlate statistically product quality with some analytical equipment raw data.
So if such work, on modeling, is developed and then transferred to the shop floor, it’s enabling to have inline real-time quality monitoring in the process. And this is as a first perspective. First, I mentioned that it is a huge boost to performance because it eliminates all the manual work that we are used to seeing in the industry on managing samples between production facilities and quality control lab facilities.
And all the necessary steps that are also eliminated and the time and the cost associated to it from the standpoint of extracting the sample to the moment where the sample is actually analyzed and results are produced, checked, and then eventually transferred back to production records that will potentially drive decisions to make process adjustment.
All these sequence, all these cycle takes time, involves lots of labor from different key roles in operations. So comes at the cost that is significant. And because it’s so manual, it has inherently some variability. All of that is now streamlined by using PAT and so productivity, consistency, batch-to-batch, and consistent quality.
Also, of course, compliance, as I said, comes to my mind. So less transcription of data. Less variability. Also the fact that there is less variability allows to operate closer to quality limits and that’s why I talk about throughput as well. So you can challenge the process further. And safety I’ll leave it to last because there’s I would say minor contribution to safety because there’s less exposure of people to the product since we are eliminating all the sample management.
Regarding what industries can it fit? Yeah, definitely Life Sciences, which is where I’m coming from and it’s closer to my heart, but definitely any process that has chemical transformation involved and requires chemical analysis. And so that will span from Life Sciences through oil and gas, chemicals, pulp and paper and metals and mining, I would say.
Oh, and by the way, food and beverage definitely as well. So let me summarize these. Life Sciences, oil & gas, chemicals, pulp & paper, metals & mining, and food & beverage.
Jim: Which pretty much sounds like most of the process industries and hybrid industries and yeah, so if there’s sampling involved in that, I think you really hit on those challenges really well of, you know, it’s manual, so it’s time-based, there’s time lags and getting the information back to make whatever adjustments need to be made, a whole host of things that if you can automate that process much more, it’ll drive a lot of really good things from quality to availability, to throughput that you mentioned.
So let me stick with you a minute here, Jorge, what strategic and operational aspects should an organization consider in introducing?
Jorge: So as in all things that are related to digital transformation, I would start by people reviewing work processes and finally the technology. So starting by people as a standpoint of strategic actions, it’s very important to assess if the organization has the right people. And so I’m talking about PAT scientists, data scientists are key for the correct preparation of introduction of such kind of technology.
In terms of the actual work process or strategy, it’s very important to consider first and foremost, to prepare in the development space, the introduction of PAT, but always, in my opinion, with a mindset of bringing over to manufacturing operations. And the reason why I insist on this is because there is plenty ways to do PAT and there are plenty mathematical packages or strategic statistical packages, if you will, and PAT platforms that are products that come off the shelf in the market, not all of them have the best features for operations.
And namely the level of customization, the level of effort in software validation comes to my mind. So it’s very important to have comprehensive, deep analysis of what’s the solutions out there and how the organization wants to fit them and where wants to fit them for their particular use case.
And finally, in terms of the operational aspects, well the obvious one is to consider what are the organization’s products, quality parameters, what are the relevant quality parameters? And make the choice of the PAT instruments that are required, get them in place, integrate them with the production assets.
As a last point, to consider is the integration with not only the control systems but even with manufacturing execution systems to reach the peak of value, the maximum value of the technology, which would be getting to real-time release.
Jim: Yeah, it sounds like, you know, oftentimes people want to start with the technology, but what you just described is really starting with the people and work processes and then working your way forward with what you can do and really looking at something built for purpose seems like a very important thing there.
Bruce, let me. Turn it over to you. What is the technology behind DeltaV Spectral PAT?
Bruce: So, you know, as, as Jorge mentioned that there’s a lot of folks that are trying to put this together. And with the advent of these really smart analyzers, spectral analyzers, we’ve seen a lot of these deployed in the process, but it’s a very complex architecture, because there’s computers that are up at the L3 portion of the network, there’s the L2 portion, the data exchanges between the disparate systems.
And so, when we looked at all that, we said, well, what could we do to collapse that architecture and make it easier to deploy? And, that’s how we evolved into our DeltaV Spectral PAT.
So, we now bring the spectral analyzers directly into DeltaV. We bring in the chemometric models that the scientists have developed also into DeltaV. We execute the models then against the spectral array directly in an Application Station on the DeltaV side, and produce the predicted results. So, now we don’t have to worry about multiple computers, multiple validation touchpoints, fragileness of the architecture, because it’s all solidly done at the DeltaV level.
Jim: Well, yeah, that sounds like it simplifies it. Just trying to move data around and make sure it’s in the right format and in a timely way. The network connections are good between sounds like, it’s simplified quite a bit. So Bruce, I guess what PAT instruments can it handle?
Bruce: That’s a great question. Cause that was one of the focuses that we had when we developed a DeltaV Spectral PAT was to make use of OPC UA. And so really any spectral analyzer that can put out its spectral array as an OPC UA floating point array we can consume with DeltaV. So we are not at all specific to the analyzer that’s used.
We can talk to any analyzer that supports OPC UA.
Jim: Well, that’s good. So as long as the supplier has it to the standard of OPC UA, then we can bring it in and process it and do all the stuff we can do with it. Jorge, there are various PAT software platforms in the market. What’s unique or what is the unique value of this solution to the industry?
Jorge: Yeah, so this question is quite linked or to what Bruce said when he introduced DeltaV Spectral PAT solution. It’s an embedded solution. So naturally more simple, more robust, and that enables automated process adjustments and end-point decisions. Both of these have additional value to our customers. And so these are generating a higher return on investment and less risk to operations.
Jim: Interesting. And I guess with AspenTech being our partner and one of the PAT software vendors, is there value or does it make sense in adopting both Aspen Process Pulse and DeltaV Spectral PAT?
Jorge: Absolutely. Absolutely. I normally say that these two products are independently producing value to the end user in different ways for different purposes. I would say Aspen Process Pulse, it’s all about predicting the near future.
So when you’re running the operation using Process Pulse, you have an additional monitoring capability that’s giving you a near future prediction of your process trajectory using a statistical model based on a past data on a number of batches or segments of data if you’re talking about continuous manufacturing it allows for that model to be generated and plotted in, in an operator display, and then projects the current operating conditions versus that representation of the population of the past data that has been made available.
And so it’s allowing the operators to have an additional pair of eyes, let’s say, to help to detect or predict any kind of excursion to quality. and take preventive measures to avoid any process deviation.
Spectrum PAT is more about the hear and now. It’s about the real-time monitoring and immediate response to operating conditions and enables closing the loop between the quality control and the critical process parameters control.
Jim: Well, that does make it sound that they’re complementary and can provide value in doing that.
Bruce, I guess back to the spectral PAT technology, what are the solution requirements and limits?
Bruce: So, we have launched our DeltaV Spectral PAT now with version 15. 15.LTS of DeltaV, and we support both the Sitorius SIMCA Modeling application along with AspenTech’s Unscrambler application for generating the models.
And then, as I mentioned, we would then transfer those models to an Application Station on the DeltaV side, where we would also tie off the spectral analyzers and then using standard function blocks in DeltaV, convert or process that spectral array against the model that we’ve transferred in, in order to generate the desired value.
So, it runs in an application station, version 15 and above of DeltaV, and then from a licensing perspective, it’s done on a per-function block basis. So for each spectral PAT function block, we would license that to run on that Application Station.
Jim: Okay, that sounds easy enough. So for a listener that’s been listening to us for this long and wants to see it in action, can it be demonstrated?
Bruce: Yeah, Jim, we’ve got a couple of things, ways we can demonstrate. So first off, a plug for our new Life Sciences showroom that we have here in our headquarters in Round Rock. We have a Raman Analyzer connected to a two liter bioreactor where we’re measuring the glucose in the bioreactor. We can do that live in front of folks and show how the models can run against the spectral array.
We also have a very similar setup where we’ve got a virtual ramen analyzer and do all of this as a virtual session where you can see the models run against the spectral arrays and even do closed loop control back into DeltaV on the example is a perfusion bioreactor.
Jim: Well, that seems like, something.
Make the trip here to the Austin area to take a look at that. That sounds…
Bruce: Absolutely.
Jim: And Jorge, I guess from a financial perspective, do you have any recommendation to justify such an investment?
Jorge: Yeah, absolutely. So I would suggest to start by defining a strategy, which means start by figuring out what are the company’s key goals?
What are the operational pains to eliminate or to mitigate? What are the targets for affected metrics that are expected to change? Estimate costs, estimate the benefits of the selected solutions, and finally, calculate the return on investment based on the quantification of costs, but also the benefits. And finally, in terms of an approach of a proof of concept.
Yeah, definitely we recommend to start by, a sandbox environment, that you can use as platform to learn, to develop standards and to validate such standards, generating the templates and all the associated necessary applicable compliance documentation. And then use that to instantiate to create any necessary replicas that will be then rolled out to production life systems.
Jim: Yeah, that proof of concept is an important step in the process. After you’ve looked at all the benefits that you can deliver and do that ROI calculation on it. And I guess to wrap things up, Bruce, where can our listeners go to find more information on this important topic of PAT?
Bruce: So you can head right over to the Emerson.com website. We have a dedicated page for DeltaV Spectral PAT on the Emerson.com site. There’s, several videos that you can watch. We’ve got our product data sheet available there. There’s also an eBook that can be downloaded as well as connecting with the proper sales office to get more information and be able to kick the tires with our offering.
Jim: Well, that sounds straightforward enough. And I’ll add a link in the transcript on exactly where the page is. Well, Bruce, Jorge, I want to thank you both so much for joining us today and sharing your expertise with our listeners. So thank you.
Bruce: Thank you. Appreciate the opportunity.
Jorge: Thank you. Same here. It was a pleasure.
