Closed-loop validation leads to more reliable machines, exceeds customer expectations: Podcast

Bill Butcher: Welcome to Siemens digital industry software podcast series on intelligence performance engineering, brought to you by the Siemens thought leadership team. As we’ve learned throughout this podcast series, industrial machinery manufacturers are seeing an increase in complexity associated with the flexibility needed to respond quickly to consumer demands.

During this podcast series, we’ve been discussing how machine builders can leverage the intelligence performance engineering solution as one way to address this complexity and also increase flexibility. In our 4th and final podcast of the series, we’re discussing another differentiating theme of the intelligence performance engineering solution – closed-loop validation. Closed-loop validation allows the machine builders to validate their simulation by capturing and testing the relationship between the requirements virtually.

Throughout this entire series we’ve been fortunate to have Rahul Garg and Giulio Camauli with us. They will also be able to join us one more time today, and as you may recall, Rahul is the industry leader or the industrial machinery industry and Giulio Camauli is the industrial machinery industry leader for the simulation and test solutions within Siemens Digital Industries. Gentlemen, thank you and welcome again. In first three podcasts, we talked about the evolution of technology within the machinery industry and the key trends that continue to shape the industry today.

We also introduced the intelligence performance engineering solution at a high level where you guys both discussed the three key differentiators: multiphysics simulation, integrated design and simulation, and closed-loop validation. In our last two sessions, we took a deep dive into multiphysics simulation and integrated design and simulation.
In this session, I’d like to close out our discussion on the differentiators for the intelligent performance engineering as we look at a deeper dive on closed-loop validation and why this is important. Let’s have that be our first question, Rahul. What exactly is closed-loop validation and why should it be important for machine builders and their suppliers?

Rahul Garg: Thanks, Bill. This is probably one of the most exciting aspects of what we are trying to help our customers with – this whole concept of closed-loop validation. So when you look at how simulation is done, one of the things that happens during that whole process is there is a simplification of reality to ensure that you can at least get the simulation contest done. And that’s part of their whole simplification process that happens is you miss out on the true criterias and the true environments under which the machine could be working.
So to make sure you do not miss out on any of the those assumptions, you make sure you do not oversimplify your design and your upfront requirements. It’s critical that you’re able to test and validate your machines in a real environment. And that’s where the closed-loop validation process comes in, where through intensive scrutiny of how the machine is performing, you can get a real sense of how it is meeting the customers’ needs, from production rates, from scraps, from noise, from vibration, you can get a real time feedback, and then you can correlate that to your simulation environment.

What this allows you to do is to make sure that your simulation models are accurate, and you are able to address even potential issues before they come out in the field. And you’ll be able to evaluate how your machine is performing on an ongoing basis. And this really helps you evolve your machines to the next generation in a very rapid manner.

Bill Butcher: So you mentioned the excitement that you have around this topic, which I think is fantastic. And I think for machine builders or machine manufacturers, it might be a little daunting, right? I mean, they’re hearing about connectivity. How do I do that? What does that mean for me and my processes? So what current challenges are machine builders facing around the topics of data connectivity and test based validation today?

Rahul Garg: So when we think about the issues around data connectivity and test based validation, getting your machines harnessed and wired up to with the appropriate sensors through design and simulation processes is one aspect of it so you can actually get the necessary information from the real operation of the machine with the right connectivity using 4G/5G connections. The bigger issue that comes about is many times today from a business perspective, the end customer who’s using the machine may not be too keen to share the data that is coming out of the usage of the machine. So those are some complexities from a business perspective that are coming in. As part of that process, your ability to work, obviously, with your end customer, so that they can also realize the benefit of sharing that data with you, obviously becomes critical. And then obviously, there are many, many more ways to do that test based validation. Giulio, perhaps you can comment on some of that?

Giulio Camauli: Yes, the fact that it’s very important that you can measure from the field, so the real usage information are fundamental for many different reasons. Knowing how the machine is used can help the designer to improve the product. But it also gives an important help to the end user because knowing the real usage of the machine, you can anticipate maintenance before going into a critical downtime issue. The digital twin is also very important and can bring value at this level.

Just combining test and simulation, you can acquire a number of channels through physical testing and physical points, you can measure acceleration, forces, temperature, noise, whatever. But sometimes it’s not so easy to instrument your machine, so you can integrate and complement the physical measurement with virtual measurement that comes from virtual sensors. So you having the digital twin of the machine, you can virtually acquire data from your model and sync with your real measurement. That’s a big advantage of sharing the information between test environment and simulation model. So having the ability to import your simulation model into the test environment makes very easy to have a huge set of information all together synchronized and verify the real condition, combining measurement with animation, with videos, with sounds, and all this information will be really helpful.

Bill Butcher: So Giulio, taking that little bit further, you know how can intelligent performance engineering solve some of the challenges that machine builders are facing specifically in the context of validating machine simulations and validating the relationships between these machine requirements?

Giulio Camauli: Yeah. That’s… Starting from requirements – that’s interesting because sometimes you have in mind a competitive solution as a market target. But how can you verify the performance of that machine? You don’t have the design. You don’t have the simulation model. The only thing you can do is to measure a certain performance. Imagine that you want to reach a noise level that you like in your competitor product. You can measure those values and use those measurements in your simulation model to verify which solution you need to apply to your design to reach that objective. That’s the first use of test to help simulation. Another easy mechanism where tests can help simulation where you want to verify the realism of your simulation model. So simulation is fine. Simulation is really saving a lot of time and cost, but of course, an error in simulation can bring a lot of damage. So sometimes you need to do some measurement. You don’t need the complete prototype. Maybe you can use a measurement that you already have from a predecessor, or you can measure just a part or a mechanism just to verify if your simulation model is representing the reality.

Then, consider again the combination of test and simulation if you needed to certify your product against a standard – acoustic standard, vibration standard. It is very important, of course in this condition, you can’t avoid the prototype. The certification need to be performed on a physical object, but if you have a digital twin of your product, you can simulate and preassess the risk in your final acceptance test.

Working on the medical model with this way, you can go, let’s say you can pass your certification immediately the first time that you go there. It is a damage if you will not pass the certification, and you have to restart from design to apply modification to correct the problem. It’s much better to reach that front on the simulation model and go to the certification phase in a smoother way.

Bill Butcher: So I wanna quickly go back to something that that you mentioned a little bit earlier here in the discussion. It’s the capture of real-time data from the field to help validate the virtual simulation. You mentioned about that and service opportunities, things like that. Can you take that a little bit further for our listeners? And and how does this happen exactly and why is it so important for machine builders?

Giulio Camauli: The connection between the simulation model and real life and the machine in the real-life condition has two important effects. The first one is that measuring the behavior of the machine when it’s used by the customer, by the end user, it brings a lot of value to the design process and to the manufacturing process. So having the ability to collect data in a centralized and safe environment can bring a lot of information that you can use that as an input for the digital twin that can predict different conditions under those real loads and real missions.

So you can verify how the machine is evolving and how the machine is behaving in order to plan your maintenance, avoiding the critical condition of finding a suspected problem – a downtime situation. But I started from the end. It is also important if we consider that having this digital twin is also very important when you deliver the machine.

When you bring your machine on site, when you deliver that to your customer, and you need to commission it. So again, one of the objectives is to get an acceptance as quick as possible without any problem. And you can do that, again, anticipating the verification and the optimization process. For instance, so if your PLC code for the automation of the machine just working with real PLC, the physical controllers applied on the digital twin of the machine. This is nice because you can then improve your code, having in mind the objective like, again energy efficiency while you increase the throughput rate. And when you go on site, the commissioning supported by the virtual commissioning will be really, really easy.

Rahul Garg: To me, this whole concept of bringing in the simulation environment in the test environment takes it to the next level where you are thinking about how to make your machines more adaptable. Right, so as you’re getting the real feed from your machine performance by applying that against your different simulation environments, you can determine if you need to make any changes to the actual performing, to the actual usage of the machine. So that’s where the adaptable nature of this whole concept comes in.

Whereas, say for example, you see that by running this one motor at that 100 RPM, you’re getting some throughput issues. Perhaps you could use that information to make it adaptable and say, you know what, if you were to run it at 70 RPM, you will still get the same throughput and it will not have an impact on the overall performance of the machine, and we will not run into any heat issues at least at the moment. So that’s the adaptability of being able to obviously address the issues, but also being able to do those issue resolutions in a real time basis. Not only just issue resolution but even trying to improve things as they are going. So to me, I think that’s the other very interesting part about this concept of closed-loop validation.

Giulio Camauli: And another important point is when you try to improve the throughput speed that is a typical requirements from the customers, they expect that you can maintain, if not exceeding the accuracy. And you know, when you increase the speed, the typical problem is that the machine will start vibrating. Speed generates vibration. Vibration means less accuracy. So the capability to predict all those behavior in the simulation model but having a proof of the reality with the physical information coming from the field will help a lot.

Bill Butcher: Rahul and Giulio, I can’t thank you enough for spending the time with us discussing the unique value delivered through the intelligence performance engineering solution. Throughout this series, we covered a lot of topics, and you guys delivered incredible insights into the challenges and trends that are driving the machinery industry today. And before we end the podcast series, Rahul, I have just a couple more questions for you. From your perspective, could you quickly summarize why or what are the main advantages for the machinery industry to embrace this digitalization topic?

Rahul Garg: To me, there are a few key important things that come out, right? The first one for intelligent performance engineering is you’re making your engineering a lot more efficient. You are making a lot more easier and faster to get your products out the door – number one. Number two, you’re reducing your risks. You’re being able to test and evaluate and determine alternatives all possible scenarios, and the more you can do, the more you can test, the smaller the risks that you may encounter in the field. Then the third piece is you have a managed environment which helps you speed up your process, which helps you serve demanding customer needs of multiple configurations, multiple changes, and you can very quickly respond to those customer needs. To me, those are the three big advantages that one can see by embracing intelligence performance engineering.

Bill Butcher: One last question before I let you go. If I’m a machine manufacturer today, why do you feel it’s critical that I evaluate the solution we’ve been discussing? You know, why should the intelligent performance engineering solution be considered?

Rahul Garg: The whole idea of the intelligent performance engineering is kind of, I guess, summarize in three key areas. One, number one is being able to do multiphysics simulations. In the real world today, there are many different issues that are faced by our equipment. Being able to evaluate and test all those different situations simultaneously under one umbrella becomes very critical – all the way from thermal analysis to fluid analysis to vibration analysis, electromagnetic interferences. All these will keep rising, right? So being able to do that simultaneously makes it really easy. So that’s the whole multiphysics simulation.

The second is the integrated design and simulation where you have one environment to bring your designers, your simulation experts all together, your suppliers together into that whole process. So it gives you the ability to quickly respond to customer changes to customer needs and to varying configuration needs and do your testing and design iterations a lot more faster, evaluate a lot more possible scenarios a lot more faster.

And then the third piece is the whole closed-loop validation. You are able to bring in real life situations into your design engineering environment. This is where the industry is going to turn a corner in a big way where you can bring in that real life information and use that to help address customer issues when they’re using the machine and help you improve your overall performance and overall design, overall capabilities of your next generation machine. So those are the three key areas I think are really important for customers to consider to help improve their engineering processes and take it to the next level.

Bill Butcher: And I think it’s becoming clear that manufacturers are looking for help, right, to address the problems we talked about, and I think that you just listed three great examples of how to do that. Gentlemen thank you again. It’s been a pleasure to speak with both of you throughout this series on intelligence performance engineering. You know, we highlighted the effects – positive effects of the solution for the machinery industry, but we also gained some really, really valuable insights from your experiences. So, I’d like to say thank you again very, very much for taking the time. We look forward to hearing your contributions on additional podcast series coming soon. Thank you everyone.

Rahul Garg: Thank you, Giulio. Wonderful talking to you.

Giulio Camauli: Thank you.