Digital Twin of Project Production Systems

[00:00:00] Gary Fischer, PE: So now we’re going to move into what is going to be an equally interesting segment around the use of digital twins in the project production system. So Roberto, can you turn on your camera and video and I’ll just turn it over to you. You need no introduction at this point, so take it away.

[00:00:21] Roberto J. Arbulu: Thank you so much, Gary. So can you see me okay, Gary? Yes, we can. We can see. Okay, great. So this session, we are going to be together for an hour and 10 minutes, approximately, and it focuses on digital twins of project production systems. And the way we are going to approach the session is we have three parts.

[00:00:48] Roberto J. Arbulu: I just want to give the audience a bit of an idea how we’re going to do this. Three parts. I want to take a few minutes to frame things around the idea of video twins. I will share my screen in one minute for that. And then we have two applications, one that is focused on oil and gas. Fuel development is going to be led by James Choo.

[00:01:15] Roberto J. Arbulu: So we’re going to invite James later to join us back. And then it’s going to be followed by McKinsey & Company, who we have a couple of guests from McKinsey, who are going to share with us what they have been doing in terms of enabling digital twins of product production systems and particularly two applications that they have actually done on civil infrastructure.

[00:01:42] Roberto J. Arbulu: After that, we will spend some time on Q&A, questions and answers for both presentations combined. And to finish this session, we have a couple of additional guests that are coming from two universities, and we have Cal Poly and UC Berkeley – a couple of students that will share their work in terms of modeling production systems and contributing to these idea of digital twins.

[00:02:19] Roberto J. Arbulu: Okay. If we have time, obviously at the end we will have some Q&A for Cal Poly and UC Berkeley, but we’ll have to check on time at that point. That being said, let me share my screen and position a bit more. What we are going to discuss, and this is a hundred percent about context on digital twins of production systems.

[00:02:46] Roberto J. Arbulu: So a bit of a story. So a couple of weeks ago, two or three weeks ago, I had a chance to have a conversation with Martin Fisher, who by the way, is following after this session on an interesting presentation on the project of the future, and Martin and I were talking about the twins. And out of the conversation came this question of a twin of what, right.

[00:03:09] Roberto J. Arbulu: And what we see. We, once again, we were talking, having a discussion, and we were talking about the use of the concept of a digital twin for existing facilities, existing facilities that are operating. It could be a refinery, it could be a plant, right? Where you want to have a digital representation of the plant and the digital component, the different components in a system, and somehow get obviously a connection between the real asset that is an operating asset and this digital asset.

[00:03:51] Roberto J. Arbulu: Okay? And so that was one element of the discussion, right? But the focus today is not on that. Then we move from an asset that is operational to an asset that is being built, which is basically what we do in projects. Right? And so, we ended up talking about the use of 3D models, right? A lot of people these days, and I’m bringing this up because a lot of people these days in the industry are referring to the 3D models of the asset being built as a digital twin of the asset.

[00:04:29] Roberto J. Arbulu: Right? And 3D models have been around for several years, right? Decades. So it’s nothing new, right? I think the term is now being connected to the asset, but today, once again, we will concentrate on a very specific application of what we call here the Institute Digital Twins, and is a digital twin of the actual production system required to produce the asset required to engineer and build the asset.

[00:05:02] Roberto J. Arbulu: On this side, on the left side, we have the actual real thing being built, and on the right side we have a digital representation of not only one, but multiple production systems. And the answer is yes. In order to do this, we need to have, as you can see on the right, a way of modeling, mapping, and modeling, and digitally representing these production systems, right?

[00:05:34] Roberto J. Arbulu: Because the production systems are on the left, physically occurring, physically being put together, but obviously we don’t have the luxury and the complexity in construction. This could be deployments, too. Back to the previous discussion, it applies to both construction and deployment. We don’t have, it’s so complex that we don’t have the means in a conventional way, but these days, technology is enabling us to really get the concept of the digital twin and make it a reality digital twin of the production system.

[00:06:06] Roberto J. Arbulu: So in essence, what we’re discussing here, or what you’ll see, is a digital representation of one or multiple production systems. The actual production system happening for a project I repeat for a project, right? And a link that is actually a real-time link back to the model and the model and the detail twin providing an analysis automatically using operation science, for instance, and determining what should the optimal behavior be of the production systems.

[00:06:42] Roberto J. Arbulu: That are either, this could also, by the way, I should have said that this could be at the beginning before we have an operating process, or it could be during the operation of a given production process. Okay. So the concept is this linkage, the real-time link to the ditto twin between the real asset and the ditto asset.

[00:07:07] Roberto J. Arbulu: That enabled us to get access to a variety of information and data and insights about the performance, about what the performance should be versus what it is from a production perspective. This is not about from a conventional project management perspective. Right. Those curves that you see on the left, these are not s curves doing progress management, but it’s about throughput.

[00:07:35] Roberto J. Arbulu: It’s about understanding capacity, utilization, understanding cycle times and working process, and many other factors that are typically associated with production processes. And so that is one concept. Now we will take it to the next level. Right. And this is part of the framing that I wanted to do for our audience today, is that because these days we have access to a variety of technology that enable us to capture data about the current performance and transfer it, we can now talk about a vision called “intelligent production,” right?

[00:08:12] Roberto J. Arbulu: Where we can have production systems that automatically optimize themselves. As long as we have this connection between the virtual and the real production system, right? Using Operations Science for instance, right now, what I wanted to do here very briefly, this is actually from the Internet, there’s nothing to hide.

[00:08:37] Roberto J. Arbulu: Companies like Caterpillar are already using a series of technologies for autonomous vehicles on mines around the world, right? And the reason why I’m bringing this up is because this is happening as we speak, number one, and the use of this technology connecting to a specific resource like this truck on the screen, and that monitoring the movements of the track and how much load, for instance, it might actually carry from point A to point B.

[00:09:09] Roberto J. Arbulu: All this data being transferred to a digital twin of the actual production system. This could be technology attached to pipes, could be technology attached to like the Internet of things. And then we are closer and closer and closer, as Gary mentioned in his opening remarks, to the idea of a digital twin of project production systems.

[00:09:35] Roberto J. Arbulu: At the end, we’re talking about having a model, something that can, we can visualize what that production system is and once again, use technology as a means to connect it and have a virtual digital twin of the production system. Okay, so I just wanted to frame things a little bit because we are going to talk about multiple applications, including the students coming from Cal Poly and UC Berkeley on what is being done about this.

[00:10:06] Roberto J. Arbulu: Okay? So I will stop sharing my screen at the moment, and I will invite James to share video and enable audio please. And, James, I’m not going to introduce you as you have been in previous sessions, so let’s keep it really simple. And James will take us through an application of what others are doing in oil and gas field development.

[00:10:32] Roberto J. Arbulu: So, James, over to you. All right.

[00:10:39] H.J. James Choo, PhD: Thank you, all right, so we’re going to actually talk about the role of the digital twin in onshore field development. As many of you are familiar with, onshore field development actually is the deployment, again, the word “deployment,” deployment of something that looks like this for your gas or oil gen production.

[00:11:05] H.J. James Choo, PhD: Okay. Now, traditionally when we talked about onshore field development, we are talking about the actual drilling work or the actual production once the wells are completed. And the digital twin is very much focused on the 3D modeling aspect of that work. And so you might actually have digital twins, as many actually refer to it as Roberto already talked about and described.

[00:11:37] H.J. James Choo, PhD: What we actually mean by digital twin is the actual production process that takes the wells and the pads from the planning perspective, planning activity to the actual point of being put online. And this is a very, very simplified view, but regardless of who you are, that’s in the process of actually developing onshore unconventional will.

[00:12:02] H.J. James Choo, PhD: This is pretty much the process that everybody actually goes through. Okay. Again, there could be, this could be broken into, you know, three different steps depending on the different regs you use. You can actually have multiple completion processes here, multiple hookups. And then the main pipeline actually is a separate item that gets connected to the site.

[00:12:25] H.J. James Choo, PhD: But generally this is the process that all the onshore field development organizations actually go through. Okay. So we can actually say that’s the model of the production system. Now, one of the benefits actually of being able to make that model is to be able to understand how we want that production system to

[00:12:48] H.J. James Choo, PhD: use that information from the actual, the actual information of the, how the work is actually getting done to inform the model. So this actually can be a living system. So, for example, if you actually know what this duration, the capacity, the resources that you’re going to actually have, then based on the number of wells you’re going to try to put on online that year or the next year, the future years, you can actually determine the web targets and stock targets and the required number of resources to actually achieve that output that you’re actually looking for.

[00:13:28] H.J. James Choo, PhD: That becomes something that you’re actually using to control whether you’re actually ahead or behind in terms of your web, not in terms of time, but actually above or below in terms of your web and stocks. Okay, now, this is something that we actually all talk about a lot, is that there are many, many organizations and the systems out there that actually focus on visibility.

[00:13:53] H.J. James Choo, PhD: And the visibility actually says how much, how many pads and wells do we actually have in permitting? How much do we actually have permitted? How much do we actually have in built? Again, the first step is to actually differentiate a pad that’s rather than not actually responsible, you know, over to construction, but whether it’s actually waiting to be built or actually built and actually waiting to be drilled or whether it’s being built.

[00:14:21] H.J. James Choo, PhD: So differentiating the three status, that would be very critical. But even at that point, once you actually see that there really is not much action you can take if you don’t know what the web targets are. Okay, so, for example, let’s actually say you actually have 20 permitted wells, is that enough or is that not enough?

[00:14:44] H.J. James Choo, PhD: Okay. Unless you understand how, about what the targets are for the overall behavior, optimal behavior of the system, you actually don’t know what management actions you need to take. Therefore, setting this target by using something like this is critical. As the work is getting done, you now actually have the actual duration that can go back, feed into the model at the same time the amount of rework that’s actually being performed.

[00:15:10] H.J. James Choo, PhD: Why is the rework actually important? Because in order for you to actually get the amount of wells you actually need outside to come out, if there’s rework, then you need to actually have more work or more capacity dedicated to individual operations or actually have more in web. So therefore this is critical to the behavioral production system and the actual number of resources.

[00:15:35] H.J. James Choo, PhD: Again, the mismatch between the required versus actual will tell us whether that target that you’re trying to hit is feasible and, if it’s not feasible, when you can actually expect it to be done. So this is one living system that actually needs to say that can actually start from anywhere. So some organizations actually have started to,

[00:15:55] H.J. James Choo, PhD: we won’t actually start doing this because we don’t really actually have much data related to how these systems actually, how these operations actually function. Therefore, we’re going to actually start collecting the data and use that to build a model and then set the target and go from here. Some organizations have said no, we actually have a pretty good estimate of what we actually want to have come out of it, or let’s actually start with the model and then let’s actually control it.

[00:16:17] H.J. James Choo, PhD: Either way. Where you, where regardless of where you start, what you want is now op a system that actually operates in conjunction. Okay? So by doing so, by doing the model again as we talked about, you can actually understand what your capacity utilization actually is and is your bottom neck where it needs to be, and how much WIP do you actually need to have in the system?

[00:16:38] H.J. James Choo, PhD: Okay. And by actually having the control element, you’re constantly actually looking at your cycle time. Your, you know, what your, you know, mean, what your, the variance is. And the median. So that allows you to actually understand what parameters you want to actually put into your model. Okay.

[00:16:57] H.J. James Choo, PhD: Again, for each of the operations that you might be actually seeing. Now, one thing that we actually had done in the past is there was actually a very big effort in creating the model, collecting the data, and then actually feeding the model and so forth. But what we’ve been actually working on with our team is the ability to build in the actual.

[00:17:18] H.J. James Choo, PhD: The op modeling and optimization process into the tool that people are actually using to control the work, whether it be discrete and simulation or actually analytical modeling, that you can actually, you are able to actually see it from one perspective. So this is something that’s, no, that’s actually already possible.

[00:17:34] H.J. James Choo, PhD: Again, whether you’re actually going to use your setup process to do this, or you’re going to actually build a model to do this, it’s something that’s already been integrated so that you can actually have a digital twin that operates from the modeling to the control and the control to the model. Okay. With that, I’m going to hand it over back to Roberto and we, I can answer any questions you might actually have after other presentations.

[00:18:00] Roberto J. Arbulu: Excellent. James, thank you so much. And it is important that our audience, you know, for those of you who are not in oil and gas, you know, also think through. your own projects and how these examples might actually apply to your reality as we go through some of these discussions, right? If you’re on the deployments, we can say, how can these apply to deployments, right?

[00:18:25] Roberto J. Arbulu: We have multiple things going through and it’s important to be thinking through all this. So thanks, James, for your insights. We are now going to move to the second presentation by McKinsey, and so I will invite Shub and Stan to please share video and audio.

[00:18:56] Roberto J. Arbulu: Are you there? Hey, Roberto. Stan, how are you? Good, thanks gentlemen. Thank you for joining us. Yeah. And so let me, let me provide a bit of context for your presentation very, very quickly and introduce you, as well. And so McKinsey has been obviously looking at the application of digital twins for project production systems.

[00:19:28] Roberto J. Arbulu: They bring to us a story and some application examples from a couple of civil infrastructure projects. They will take us through the details in a minute. And I, you’ll see some very interesting, I will say impact, as well as how they did it. But before you share your screen and get going, let me introduce both of you gentlemen.

[00:19:55] Roberto J. Arbulu: So first we have with us Shubhraneel Mitra. Shub is McKinsey’s CapEx and Sustainability practice. He’s a member of that practice in Southeast Asia. He brings to bear his knowledge of the working across energy and material sector, especially across the EPC oil and gas rail and power customers, clients, and has advised public sector, private and multinational companies on various aspects of the business to improve business performance, including shift to cleaner technology like hydrogen and electric vehicles.

[00:20:38] Roberto J. Arbulu: On the other side, we have Stanislav Gaponenko, a project manager with McKinsey & Company based in Jakarta, Indonesia. He serves major engineering and construction and basic material clients in Southeast Asia, Russia, and the Middle East on operational transformations and capital portfolio optimization projects.

[00:21:03] Roberto J. Arbulu: Prior to McKinsey, Stanislav had extensive experience in fabrication EPC and offshore capital projects. Spending eight years with Chevron on ln projects in Australia and working offshore in the north. He earned a master’s degree in mechanical engineering from University of Workweek. So once again, welcome to both of you and on behalf of PPI, thank you for joining us today and we will keep it simple so it’s really over to over.

[00:21:38] Roberto J. Arbulu: Yeah.

[00:21:39] Shubhraneel Mitra: Thanks a lot for that, Roberto. And thanks for this opportunity to present some of our work in this symposium. I will set some of the context and Stan will take you through some of the details. What we will do today is share two examples of PPM applications. And these two examples are in very different contexts, right?

[00:22:05] Shubhraneel Mitra: One of them is an example of PPM application at a package level where we are trying to improve the productivity for piling. And the second application is at a project level where we try to analyze the project in its totality and try to see how we can accelerate the overall project and move the completion date ahead.

[00:22:32] Shubhraneel Mitra: Right. So these are the two examples that we want to share today. And it is the first time that we had the opportunity to apply PPM at a relatively large scale at an Asian company. We were able to do this with one of the largest EPC contractors in the region. It’s a diversified plant but particularly large in infrastructure.

[00:23:02] Shubhraneel Mitra: Especially has a spike in the area of building and managing tool roads. But apart from that, it does have other businesses in the process industry. Could be oil and gas, could be mining and smelting. So for this particular client, I think, they wanted to improve site productivity and look at lean construction, but they wanted something beyond traditional lean techniques.

[00:23:32] Shubhraneel Mitra: I think they have been practicing some traditional lean techniques for quite some time now. And that’s when we thought it would be a good idea to introduce project production systems. And the idea was not only to try and improve lean productivity at site, but the client also wanted to have a good understanding of their supply chain systems in terms of the construction at site.

[00:23:59] Shubhraneel Mitra: Be able to measure it, be able to quantify the improvement through lean and also develop internal benchmarks as they go along. Because the client at any one point in time would have a portfolio of approximately one to 50 projects running. And many of these projects are repeatable. So it was important for them to kind of go into a factory mode when it comes to lean improvement and site.

[00:24:25] Shubhraneel Mitra: They wanted to create these lessons learned in internal benchmarks. So improvement in one project could be translated to other projects also. That’s where we came in and there were a few projects, I think, for the client, that were not running on time. We picked two of them and started, if you can go to the next slide.

[00:24:48] Shubhraneel Mitra: We tried to look at it. Around five levers. So looking at the product design, looking at the process, looking at capacity, inventory and variability, and I think particularly important over here, which I think was very interesting for the client, was the point around variability. Because most of the time when the client used to do their planning as well as create their gang charts and understanding of the variability was lacking among others, right?

[00:25:19] Shubhraneel Mitra: So we tried to look at these five levers on two projects and also to understand the cost benefit analysis, right? So if we were to de-bottleneck a certain constraint, we had to add some resources and, if you were to add that resource, there was some cost involved and whether that made sense in the context of project acceleration,

[00:25:46] Shubhraneel Mitra: I think that was a further insight when we tried to do the analysis. With that, I think I’ll hand it over to Stan who will actually share the two real-life examples that we had the opportunity to work on.

[00:25:59] Stanislav Gaponenko: Thank you, Shub, and also from my side, thank you to the Project Production Institute for inviting us and allowing us to share these cases with the audience today.

[00:26:13] Stanislav Gaponenko: It’s good to be here. I’ll talk a little bit more about what the two applications were. So the first one on the left was for a Harbor Construction project. They’re both civil infrastructure projects. The scope was relatively simple. So I think for the Harbor Construction, we had some PSP and CSP piling capping beams, pile slabs, ducting system, some of the fenders and ballers on the marine side.

[00:26:47] Stanislav Gaponenko: But the project had some additional complexity in the sense that some of the work could only be done at certain times of the day because of the tidal waves and access to the areas. And when we started looking at it, it was rather significantly behind schedule due to various technical challenges in particular on the piling productivity.

[00:27:16] Stanislav Gaponenko: And for the second project on the right-hand side, this was actually on a package level. This was a road, a highway construction project. And we’ve looked specifically at one of the sections, the main bridge section and tried to improve the piling productivity on that.

[00:27:39] Stanislav Gaponenko: So again, rather simple civil scope, but also limitations for the road construction in terms of the times of the day when the work could be done. Because this was in the residential area, there were limitations on the noise due to piling. And despite kind of the simplicity of the scope,

[00:28:04] Stanislav Gaponenko: but due to the, and also due to that added complexity of the limitation on when the work could be done, we still managed to get some good results with the application of digital twins. So in the first case, we looked at about 20% schedule acceleration for a critical project milestone, which was halfway through the project and had significant liquidated damages associated with that.

[00:28:36] Stanislav Gaponenko: And for piling we actually saw up to 300% improvement in the cycle time. And this was over and above what we would normally call the classic lean construction levers. So I’ll just show you a little bit more on these two starting with Harbor Construction. So the two Gantt charts here show what was the baseline situation and the three milestones, milestone A, B, and the ultimate project completion.

[00:29:11] Stanislav Gaponenko: And you can see that Case was trending to a two month delay against the first milestone. And the overall project completion was also trending to about a one week delay. And by modeling the production system in a, in a digital twin all the processed steps, all the resources and also the variability levels, the performance that we saw on site,

[00:29:38] Stanislav Gaponenko: we actually managed to find a simulation where the milestone, which had penalties associated with it, could be pulled within the target completion date. And also the overall projection was accelerated about three weeks to end at two weeks ahead of schedule. And as I should have mentioned before, of course there was some cost associated with that acceleration.

[00:30:07] Stanislav Gaponenko: But the digital twin allowed us to know what resources exactly at what time were needed. And at the end of it, the net benefit was still there. So the avoidance of those penalties was actually a lot higher than the extra cost that we needed to do that acceleration, do that sprint to deliver a milestone.

[00:30:33] Stanislav Gaponenko: And this is just a representation of what the digital twin model looked like. So every one of these shapes or blocks represents a station workstation with inputs going in and the outputs coming out, whether it being piling or welding or rebar installation. So you can see that even for such a simple civil scope, that model can be quite extensive.

[00:31:01] Stanislav Gaponenko: And each of the lines represents relationships like network schedule, showing the transfer of work and once we’ve run the model we do a number of simulations. So at times, hundreds of simulations with the different levels of variability between the processes, and there’s just two snapshots with the two settings of flow variability and medium variability.

[00:31:32] Stanislav Gaponenko: And what it shows us is, it gives us a probabilistic model to tell when the project is likely to be completed with all the parameters that we enter. And this is just another view of what production manager software allows us to do. Each of these green and or orange bars represents the capacity for each of the resource types and resources here can be equipment like train or wibo, hammer, et cetera, or workers.

[00:32:10] Stanislav Gaponenko: So manpower is used to deliver welders, helpers, et cetera. And you can see that there is a difference in height between the bars, which shows the different utilization of these resources. And the goal of the analysis is to balance it out as much as possible, which is never, never possible.

[00:32:34] Stanislav Gaponenko: But the closer we get to a more even profile, the better the utilization of resources is over time. And essentially we do this in cycles to try and resolve the bottlenecks. And for this particular case, the bottleneck moved from the crane initially with the orange bar on the left-hand side to then go to molding equipment that was used for concrete in and then later to back to the crane.

[00:33:12] Stanislav Gaponenko: So it took a couple of rounds to optimize this production system to give us the outcome that I shared, and capacity, of course, does not stay constant over time. So there is an additional dimension of time, which we model. And the graph here shows two different types of resources of what it looked like from the beginning of the project until the end.

[00:33:40] Stanislav Gaponenko: And what this allows us to do is also know when the highest utilization is. So where the kind of the peaks on that graph are but also where the values are, where the lower utilization is, which allows us to demobilize the resources when they are not needed and save project cost. So it allows us to develop various delivery strategies based on the understanding of how resource utilization and capacity utilization changes all the time.

[00:34:19] Stanislav Gaponenko: And lastly, I think, kind of the summary of what the recommendation was after running the digital twin. So on the left-hand side, this is split into two types of resources, the equipment and all the machinery and the manpower. And on the left-hand side is the baseline case.

[00:34:42] Stanislav Gaponenko: This is what we started with and this is what the client was planning to run the project at, at a more or less constant basis. And we’ve actually restructured the resource allocation with an increase. So that increases here to deliver milestones. So doing a, a sort of a sprint and then ramping the resources down back to a lower level below what the baseline was for milestone B.

[00:35:14] Stanislav Gaponenko: And then it stayed more or less constant until the project completion. And that restructuring or reallocation of resources, depending on the time, depending on the milestones produced, as I mentioned, a net benefit because of the ability to deliver milestone A and avoiding the penalties that were associated that,

[00:35:39] Stanislav Gaponenko: and if I briefly go over the second case as well, so this is for the construction of a bridge. And the model here on the left-hand side actually looks a lot more simple. This is because we haven’t modeled the full project scope. We’ve just modeled the piling package. So it involves probably about two dozen steps to be able to do that.

[00:36:09] Stanislav Gaponenko: And we’ve continuously updated the model similar to how Roberto mentioned in his, in his framing conversation with the actual productivity on site, with the progress rates on site to run the model over and over again to give us a recommendation on what has to be changed for the best outcome to reduce the bottle.

[00:36:36] Stanislav Gaponenko: And in this case, out of the five levers that you mentioned, we could play with four out of five because we were already in the middle of execution for this project. We could not influence the product design or the engineering design of what was being built, but we could change the process design.

[00:36:58] Stanislav Gaponenko: Despite the simplicity of the work being done, there was still some sequence that could be modified to improve the execution capacity. So that was one of the main levers, adding resources where it was needed, including manpower and machinery. And also changing the shift pattern for some of the resources inventory.

[00:37:24] Stanislav Gaponenko: So controlling the WIP similar to the example that James shared, and of course the variability, both in terms of the external factors like weather and also the internal ones, the performance of, for example, welders and the quality of the work that they were doing. And for the different parts of the production workflow, you can see that the bottlenecks were quite different.

[00:37:53] Stanislav Gaponenko: So I think the main ones on the machinery side were for the cranes. And together there was about a month’s worth of schedule impact at play for resolving that bottleneck. But there were also similar bottlenecks on the manpower side. And the main one was on the rigors or the helpers that were installing the piles in place before they were driven into the ground with about twice scheduled impact at play here with about two months.

[00:38:36] Stanislav Gaponenko: And similar output from the software production manager where we tested different scenarios or different combinations of what could be done in terms of adding, for example, one hour to shift time for a particular type of resource or adding another crane or viro excavator or playing around with the batch size of the handoff between the different works stations.

[00:39:10] Stanislav Gaponenko: And this continuous optimization allows us to find an optimal combination of resources, shift time, and a sequence that allowed us to accelerate this project and improve the piling productivity and just the visual representation of what the cycle time looked like before and after the change.

[00:39:39] Stanislav Gaponenko: And the row on the left-hand side just shows the steps similar to what you saw with the different shapes on the model screenshot on the previous page. And the chart on the left is the actual productivity that we’ve observed in the beginning, or the baseline with about 170 minutes cycle time for piling and the overall productivity being about two piles per day installed.

[00:40:13] Stanislav Gaponenko: And we’ve applied the more classic or traditional lean construction levers which are quite simple. And that already gave us an improvement to about three piles per day or about two hours cycle time. And then with the application of digital twin, we went to the scenario on the right-hand side where we went to about five or, on good days, six piles per day.

[00:40:42] Stanislav Gaponenko: And just visually, you can see how a lot of the work in the beginning and also in the middle has been completely de-constrained through the addition of the second train, which essentially takes that work off the critical path and is the main driver for increasing the productivity. And I think the last page, just to show the summary of what the impact was and the different opportunities or different levers that were applied,

[00:41:19] Stanislav Gaponenko: and as I mentioned, they essentially fall into three categories: one was the addition of resources, whether it be in cranes or manpower, but knowing exactly when and where and how many should be added so that the cost does not increase significantly. We’ve looked at the shift pattern and the work calendar adding a couple of days for certain types of resources.

[00:41:50] Stanislav Gaponenko: And lastly, we’ve also looked at the sequence of work and the size of the batches. Of the handoffs between the different workstations and the impact for different levers is shown on the right-hand side. So I think, in summary, again, the main takeaway for us was that even in rather simple civil infrastructure projects, the application of a digital twin can certainly bring value, whether applied at a package level or at an overall project level.

[00:42:37] Stanislav Gaponenko: And I think it also allowed us to build that library of benchmarks that she mentioned in the beginning, which also then became a reference point for the client. Going forward with similar projects in the future. Thank you very much, Roberto. I think that’s what we wanted to share today.

[00:43:03] Stanislav Gaponenko: So happy to answer any questions or, if not, we can move to the next presentation. Yeah.

[00:43:09] Roberto J. Arbulu: Stanislav, thank you. Thank you very much. Shub as well. Thank you guys for joining us today and, and showing your work and what organizations can achieve through this methodology, this technology, because it’s really not just a methodology, it’s a combination of a given technical approach and the technology to get to the YouTube twins.

[00:43:29] Roberto J. Arbulu: So very, very impressive. So we invite James two to be on video and audio again with us please. And we’re going to go through some questions and answers. We don’t have a lot of time since we still have a couple of additional presenters. And so we’re going to go through a few questions really, really quick.

[00:43:52] Roberto J. Arbulu: I think this is something that, Stan, you, you just mentioned in one of your closing comments. And, James, can also provide and show the three of you your input on this, is the idea of creating a digital twin of a production system applicable only for systems with a large number of items going through, like what you show for wells.

[00:44:16] Roberto J. Arbulu: I think that probably we answered this question by an example that you presented on the hardboard, but I’ll let you guys expand on this, on these questions.

[00:44:26] Roberto J. Arbulu: Any, any preference of who goes first? Yeah.

[00:44:32] Shubhraneel Mitra: Maybe I can answer that. I think I have a very, very simple answer to that. Look, it all depends on the complexity. I agree with that because if you have a few inputs, a few elementary steps, and a limited number of activities, sometimes the improvement is very intuitive, right?

[00:44:55] Shubhraneel Mitra: People might be able to do it just in their brain, you know, or just using their common sense by visiting the ground. Where the digital twin becomes impactful is when you have a larger number of inputs, more complexity, because that’s when maybe using simple common sense or just going on the ground and observing the improvement ideas and the bottlenecks and the resources may not be immediately here.

[00:45:29] Shubhraneel Mitra: Right. So it’s not so intuitively clear what the bottlenecks or the resources are, and that’s where the digital tool helps us, right, to address the complexity, which the human mind cannot process that easily. So yes, I think the digital twin theoretically can be applied even in simple context or complex context, but it makes much more sense in a complex context.

[00:45:54] Roberto J. Arbulu: Thank you, Shubhraneel. James, any perspective from your end?

[00:45:57] H.J. James Choo, PhD: Yeah, so I think there’s two things to take a look at, whether there’s the numerous, actually, outputs versus the numerous repeated operations. So you can actually have one facility, that’s actually going through a production and that may involve 20, 30, 40 different pipe and pipe spools that are actually being installed.

[00:46:19] H.J. James Choo, PhD: At that point, you’re starting to already actually get repeated operations with conflict and capacity alignment between different operations. But I think that, so you don’t actually need a huge number of items to actually be produced in the end in order to actually benefit from the actual Operations Science.

[00:46:37] H.J. James Choo, PhD: Now, I think Shub actually had, Shub actually had an interesting perspective. Is it worthwhile actually doing a production system model? Especially in the effort? Sometimes if, you know, if we’re actually going to cook 10, 10 different, 10 burgers, you may not actually want to spend the time to actually create a production system map for it.

[00:46:55] H.J. James Choo, PhD: But if there’s feeding 50 people, which we actually have done in the past, we actually had to create a production system model to understand, you know, how do we reduce the cycle time so people don’t actually get cold burgers. So yeah, I think it’s not just actually about the, you know, what are you actually trying to achieve using the production the, the digital twin?

[00:47:15] H.J. James Choo, PhD: And when you actually look at individual projects, the focus actually is not about the standard products, repeated products, but it’s a repeated operation that determines whether you can benefit from it or not.

[00:47:26] Roberto J. Arbulu: Yeah, that’s a great point, James. And by the way, I would also like to propose something along the lines of this discussion, that one of the difficulties in projects, I think, is that, you know, we, it’s not like manufacturing, right?

[00:47:44] Roberto J. Arbulu: Where you can go to a manufacturing facility and you’re going to see the production system literally physically in front of you, right? You can see the production line in projects is different, right? Because we have components that will be active later, and the ability to see that in a digital model even when things have not yet physically occurred, right?

[00:48:06] Roberto J. Arbulu: And you are trying to influence, obviously become much more proactive. I think that’s what is one of the advantages that the engineering and construction industry should take from things like this. Right. There is one more question by the way. And this is unfortunately the last question for the three of you.

[00:48:25] Roberto J. Arbulu: It seems to be a question that has two parts. Let me start with the first one. The first one is about what other types in addition to oil and gas field development, James, that you introduced in infrastructure, what other type of production systems you think this can be applied to?

[00:48:50] H.J. James Choo, PhD: Well, we, I think you already actually hinted on it this morning; we’re actually already actually doing this in the data

[00:48:58] H.J. James Choo, PhD: business and actually on the deployment side of the business. So I think those are two right off the bat that they can actually benefit from. At the same time, this is being applied to manufacturing in, let’s say, facilities that are in the building. Things actually for the construction industry as well as the mass marketing industry.

[00:49:23] H.J. James Choo, PhD: So it’s not something that’s actually just actually for the oil and gas or the infrastructure, but any production system that actually has, that’s actually producing something that you like to actually optimize the performance of.

[00:49:37] Roberto J. Arbulu: Excellent. So any perspective stats and chip about you guys operating globally across multiple markets, so any additional thoughts on this?

[00:49:53] Shubhraneel Mitra: Yeah, so look, for me, it does not have to be oil and gas specifically. I think the concepts are pretty much applicable across different types of projects and what James said, and, you know, what we also felt while we were applying PPM in the context of those two particular projects, is the benefit is the benefit can be multiplied or scaled when the activity is repeatable, right?

[00:50:27] Shubhraneel Mitra: So essentially it’s about looking at a few specific activities, trying to improve the cycle time. And if that gets multiplied by, say a hundred times over the course of the project, that’s where the benefit comes from. So I would look at repeatable activities, which could be small, but when you improve a very small activity also by 10%, but multiply it by a hundred or thousand, it really makes a big difference to the overall project.

[00:50:57] H.J. James Choo, PhD: Hmm. Excellent. Good point. Thank you. one more, one more thing to add, Roberto, if you don’t mind, is, go ahead, that, so you can actually look at this from a project perspective. For, from an organization, you can also actually look at it from a shared supply chain perspective, right? So you might be actually using something on multiple projects where you actually are now competing with yourself because of the shared capacity in your supply network.

[00:51:22] H.J. James Choo, PhD: So by actually having a production system model of your supply network, you can actually start to understand, you know, how do we actually make sure that that specific production system or the value stream behaves the way we want it to behave?

[00:51:35] Roberto J. Arbulu: Yeah, yeah, you’re right. So we probably have one more minute.

[00:51:40] Roberto J. Arbulu: So these might be quick answers, but I want to pose the question to you guys. The question here, and let me read, let me make sure I get this right. How do you think this approach, obviously referred to these or twins, complement current project management practices?

[00:52:02] Roberto J. Arbulu: Any quick answer to that?

[00:52:07] H.J. James Choo, PhD: I, again, I’m not sure. Again, each different organization, each organization has different current project management practices, but actually there’s tech, the traditional project management PMI is a description of project management. I think this again, that actually has its role to play.

[00:52:28] H.J. James Choo, PhD: But I think PPM is something that can be actually augmented and that does not actually have to rip out traditional practice until you actually identify yourself, that there might be duplicate or actually better ways of handling things. So, yeah.

[00:52:45] Shubhraneel Mitra: Yeah, I agree with you. I think PPM goes pretty much in line with PMI or PPM is rather a sub element of PMI, but I think very, very closely aligned with, say, project management

[00:53:00] Stanislav Gaponenko: practices.

[00:53:01] Stanislav Gaponenko: Yeah.

[00:53:03] Roberto J. Arbulu: Alright. So look I really want to thank you, the three of you, on behalf of PPI, for the time. I know it’s really late in the middle of the night for you guys, for Mackenzie. And so thank you so much. We want to now move into – we were not able to answer a lot of the questions coming in, so apologize to the audience, conscious of time, but we will move to our next section.

[00:53:30] Roberto J. Arbulu: So thank you, Stan. Thank you, Shub and James. Yeah. Appreciate it. So with this being said, we like to invite Corey Nader from Cal Poly and Guillermo from UC Berkeley. Corey, how are you? Good. How about you? Excellent. Guillermo, where are you? Okay, now I can see you. So before I introduce you guys, let me position this a bit better for our audience. It is important to highlight that, you know, one of the PPI objectives, or the only one, one, a very important objective is to develop and disseminate the value of knowledge of PPM.

[00:54:17] Roberto J. Arbulu: And this is a combination of doing certification programs. Doing research and also this collaboration with universities like Cal Poly in the case of Corey and UC Berkeley in the case of Guillermo. And so there are other initiatives, and that later on we want to hear about what PPI is doing with education and institutions.

[00:54:42] Roberto J. Arbulu: But today we have the chance to collaborate. PPI had a chance to collaborate with Corey, which we will start with you, Corey. And Corey is probably a great example of a professional, she’s no longer with Cal Poly she’s working in the industry, correct? Yes. And you finish your undergraduate studies.

[00:55:07] Roberto J. Arbulu: But let me introduce you a bit more on this. So she’s currently a project engineer at Clayco. She’s a recent graduate from California Polytechnic State University in San Luis Obispo with a Bachelor of Science and Construction Management, where she completed her senior project in the area of project production management.

[00:55:28] Roberto J. Arbulu: She has worked mainly on industrial projects which range from a three point for mineral square foot fulfillment center to a 500,000 square foot manufacturing facility. And so, Corey, you had the chance to apply and learn about project production management and start modeling production systems as part of your undergraduate studies.

[00:55:51] Roberto J. Arbulu: Right. I think you are a great example of a professional that comes out of school with not only more traditional perhaps construction management, project management thinking and ideas, but also with a stronger focus on production. And so thank you for joining us today. And that being said, I’m going to basically, over to you and you have a few minutes to share what you did at Cal Poly.

[00:56:16] Roberto J. Arbulu: Okay. Okay. You might want to share your screen. Thank you.

[00:56:20] Corie Nalder: Thank you, Corey.

[00:56:26] Corie Nalder: Okay. Thank you again to PPI for allowing me to present my research. This is exciting. It was originally just my senior project at Cal Poly, but being able to present it for quite a few people, it, it’s an honor. So I appreciate it, and thank you for inviting me. So I, again, to restate, I work at Clayco as a project engineer now.

[00:56:49] Corie Nalder: I was an intern there and that’s kind of also what got me into this specific topic. But I used project production methodology to compare on site prefabrication, steel erection to traditional stick-built. So the intent was project production. Management provides continuity, predictability, and optimization of a process.

[00:57:12] Corie Nalder: In comparison to the traditional approach of project management, we generally see project management taught in schools. So it was an honor and great experience to actually be able to be taught project production management as like another class. So I could compare the normal traditional style to this new methodology and this new delivery type that is definitely being implemented.

[00:57:33] Corie Nalder: Project production or project management is constrained just by scope, budget, and schedule. Project production management incorporates operation sciences into that traditional delivery method, which enables the additional focus on demand production and variability. The objective of the research was to actually explore the benefits of project production management when implemented.

[00:57:55] Corie Nalder: Project production management can be easily applied to prefabrication. Prefabrication provides that continuity. Reduce variability and easy assembly alterations. So I was interning when I was actually introduced to a new type of style of steel erection. It was being patented by building zone industries.

[00:58:14] Corie Nalder: So if you can see in the little photo, there’s this panelizing system, or a panelizing table is what they call it, where they place joists and metal decking on top of the sys test table. They then create a system, I refer to it as a panelizing system. Then they lift it, pick it up, and transport it into the bay and put it in the bay, which reduces the amount of time their workers are at elevated heights and when actually implementing this system.

[00:58:41] Corie Nalder: So my research then focused on the capacity utilization. I wanted to compare the two methods, the traditional stick-built versus this new on site prefabrication style of steel erection. Again, with that emphasis on the capacity utilization.

[00:58:58] Corie Nalder: So my methodology, I wanted to begin with a main constant, which was I created a warehouse in Revit. The warehouse was comprised of two stories. It was 224,000 square feet and it was made of 140 bays. The bays were about 40 by 40 feet. They had six joists, and then there were 32 metal decks in each bay.

[00:59:22] Corie Nalder: And then, you know, your traditional columns as well as girders and beams. So after I had that model and I could apply it to both types of steel erection, I then interviewed BZI, so I could get their production rates as well as the durations. So one of my main things, I produced this Revit model to them, and they gave me a direct, like a duration of about 10 days to complete this large of a warehouse.

[00:59:46] Corie Nalder: And that included all the detailing, assembly items like that. They generally complete about 35,000 square feet per day in their steel erection. They average about, again, I’m going to refer to it as a panelized system, in about 40 to 45 of those on that panel panelizing table per day. And then it takes about 10 minutes to create that panel system.

[01:00:08] Corie Nalder: It takes about 3.5 to 7 minutes to transport the panel system from that panelizing table to an actual bay and then to place it. So I took that duration, those production rates, and I created in PPI’s process map, where, as you can see on the screen, the on site prefabrication and then the traditional model.

[01:00:31] Corie Nalder: And I put those production rates in both of those so it could be applied. And I ran through the system to see what would happen when these production rates were applied to both methods. And so then I could compare the capacity utilization for the crews working on it. So my results, I basically concluded that capacity utilization was reduced for the on site prefabrication in comparison to the traditional stick-built.

[01:00:59] Corie Nalder: Some major differences that we can see is that BZI displays reduced capacity utilization in the areas of transportation, joist installation, and bay erection. Differences that I want to also point out are the differences in crews it required to place the joists and metal decking on the panelizing table versus being raised at an elevated height.

[01:01:20] Corie Nalder: And it also requires different transportation styles. However, I could still apply those production rates similarly to each style and method. I would on – I also want to just kind of touch on the fact that reduced capacity utilization means that the crews are being utilized less to meet those demands of production.

[01:01:40] Corie Nalder: And so continued exploration. I’ve also reached out to BZI. They have actually referenced my research in some articles. I have that linked and I will be able to show you as well. They’re continuing this idea of looking at the assembly process and a lot of companies we’re seeing are implementing project production management to actually view the operations and systems and so they have a better idea of variability and can control that variability and provide better continuity of their product as well.

[01:02:13] Corie Nalder: So, one of examples, they implemented what’s called a wall master and it actually allows the IMP installation system to enhance safety efficiency and simplicity of the IMP wall. So again, similar to that panelizing table, it’s performed on the ground and then lifted into place. And then just to show you a little bit of their website kind of referencing the research,

[01:02:38] Corie Nalder: and they’ve also found that this is, provides, again, their workers are not at high areas and they’re not in the air for as long, so, which again, provides better safety for them overall. And then again, like I had mentioned, it reduces their capacity utilization of their crews. Corey, if

[01:03:01] Roberto J. Arbulu: you’re, if you’re sharing some, if you’re sharing the link, we’re not seeing it unless you are not.

[01:03:05] Corie Nalder: Okay. If you want, you can also, once we share the slide deck, I believe we can, it can follow the link. I don’t believe it’s a shared screen, but okay. No, fine. I can go back if

[01:03:16] Shubhraneel Mitra: you’d like.

[01:03:18] Roberto J. Arbulu: That’s fine. Yeah. Let’s do it. Let’s take a look quickly just so our audience can see the impact of your work.

[01:03:30] Roberto J. Arbulu: And Guillermo will be with you in a minute,

[01:03:33] Corie Nalder: and I believe, can you see it

[01:03:35] Roberto J. Arbulu: now? No. For some reason is it’s not coming up, but

[01:03:41] Corie Nalder: that’s okay.

[01:03:42] H.J. James Choo, PhD: The link will take you, that’s

[01:03:43] Roberto J. Arbulu: okay, to it. Thank you. Excellent. Excellent, Corey. So thank you so much and also, on behalf of PPI, for your work and interest in production management.

[01:03:55] Roberto J. Arbulu: Right. And so now we want to move to Guillermo Pradofrom UC Berkeley, and Guillermo, you are with us here on video. Yes, yes, I am. So let’s get ready with your presentation, just be conscious of time, please, and as I provide a digital introduction for yourself. So Guillermo is actually a graduate student at UC Berkeley at engineering and management school or program more specifically in civil engineering.

[01:04:32] Roberto J. Arbulu: And Guillermo has worked with public and private organizations in South America and North America in implementation of building information modeling. Guillermo is part of a technical committee that assists the Peruvian public public sector. He’s from Lima, Peru, with the development of the National Beam Policy.

[01:04:54] Roberto J. Arbulu: His current areas of interest include Beam lean production, application of operational science in Puerto Val here in a civil engineering degree from POF University del Peru, a certificate in BU designer construction from Stanford University. And currently, as we said, he is part of the engineering and project management Master of Science program at UCBerkeley.

[01:05:18] Roberto J. Arbulu: So Guillermo, welcome to the symposium and it’s over to you. You have a few minutes to tell us what you’re doing. Thank you,

[01:05:27] Guillermo Prado: Roberto, for the interaction. This for the, for your operation institute for improving this work. I’m going to present part of the research that I am developing as my master thesis, which is called “Impaction System Design on Project Performance Operations Science Analysis.”

[01:05:42] Guillermo Prado: In this case study instruction, I will show to you what were the research questions of my research and explain what production system I mapped, modeled, analyzed. And finally, I will summarize the most important outputs and conclusion of this study. The research questions are related to how we can apply operation science analysis, including the graphs and equations, to find the impact of production system design in project performance.

[01:06:11] Guillermo Prado: We are using a case study in building construction to understand how Operations Science Lens can help us to better understand this relationship between production system design and project performance. How can, what are the assumptions and under what circumstances we can develop these methods? The production system that I’m going to talk about is a cladding on site assembly of an offsite production system.

[01:06:37] Guillermo Prado: We are going to talk about, since the components of the CLAINE system arrive to the project site, how the EIFS panels are installed on the building elevation, how the windows are assembled throughout, through the openings of these panels, and finally, how these components of the cladding systems are caught in the building elevation.

[01:06:59] Guillermo Prado: As you can see in this slide, the process that I map is divided into entries. Principle chunks of work. The first one is related to the EIFS panels, installation and inspection. Then the windows, installation and inspection. And finally, the development and inspection of the cock. I’m focusing on understanding the demand under these production systems and how it behaves under these circumstances.

[01:07:22] Guillermo Prado: I use a demand of 16 EIFS panels per day to consult. It’s important to mention that these are engineered to order elements, and this map has been enabled by SPS project, production, optimizer and simulator. An important part of the process mapping is to understand how to collect data in this case, data related to production parameters.

[01:07:46] Guillermo Prado: Throughout the different interviews and meetings that I have with the project team of this construction process in Northern California, I collected the project plans by using the five levers of project systems optimization by PBI. I try to understand how, what are the assumptions that I have to make in order to provide reasonable assumptions to understand what are the transfer batch, the process, batch demands of each of the items that are flowing throughout the operation system, as well as the process rates and process times of the different steps involved in this production system.

[01:08:26] Guillermo Prado: These two tables represent the data and the input that I use for developing these production systems. Then we have the outputs under the demand of 16 EIFS panels per day to be installed. As we can see in the first graph, the relationship between operations and its metrics under lead flow are explained by throughput, working process and cycle type.

[01:08:51] Guillermo Prado: This green horizontal line represents the demand of 16 pounds per day, and these other green lines in the, in the upper part of the graph represents the maximum throughput. As we can see, the demand is below the maximum throughput. The production system can respond positively to the current demand.

[01:09:10] Guillermo Prado: So you can see as well, in the second graph, the capacity utilization of each of the resources. The bottleneck in this case is the last resources with capacity utilization of 83% approximately, and we can understand these metrics better in this table below, which shows the operations and metrics working process, triple digitalization and cycle time, and how that relates with project management metrics, in this case duration.

[01:09:38] Guillermo Prado: I also included the real duration of the different process included in the production system, which is this last column that validates the results of the models in those two pieces of information are

[01:09:50] Roberto J. Arbulu: similar to each other,

[01:09:53] Guillermo Prado: but I then change the demand to something larger. I enlarged the demand and instead of 16 panels per day, I use 20 panels per day to see how the production system is behaving under these new circumstances.

[01:10:09] Guillermo Prado: As we can see in the first graph, the demand now is above the maximum throughput line, which means that actually the operation system as, as it was designed before, cannot perform positively under these new circumstances. This can also be understood by the second graph, which shows red line under 100%, and two of the resources used for this production systems are going above that, which as we can see from, as, as we can understand from Kingman equation, when resources going closer to 100% or above, that we are going to see cycle times that are very large, which is the result that I am showing in this table below, in which the working process and cycle time are going to infinity because the utilization is larger than 100%.

[01:11:01] Roberto J. Arbulu: Between these two

[01:11:01] Guillermo Prado: results, I conducted sensitivity analysis of the demand I developed. 12 runs changing. The demand of all the items of this production system shall understand how the demand makes changes in Operations Science metrics and project management metrics. This demand, these results are shown here.

[01:11:22] Guillermo Prado: Part of them, one important outcome of this analysis is to see how the bottleneck, which are those orange cells, change from one run to another, which means that it’s not always the same product flow that causes bottlenecks on the production system. We can understand that by seeing how the orange boxes are changing from each of, in each of the runs of the sensitivity analysis.

[01:11:47] Guillermo Prado: We can see on the second graph how the different operations and these metrics are changing by a different demand, which help us to understand how we can positively impact these metrics and therefore have a better project performance. Finally, the main conclusions of this study is that actually we can collect data by making reasonable assumptions from the project side to understand the project, the production parameters that we have to include on our production systems models, the operation science analysis equations.

[01:12:20] Guillermo Prado: The graphs therefore are applicable to any type of construction process. We conducted one case study, but we believe that this can be expanded to more production systems in construction and any change in the demand because of different behaviors in the production system, specifically in neutralization and cycle time.

[01:12:38] Guillermo Prado: If we want to perform positively under these new circumstances, such as the enlargement of the demand, we will require some optimization and adjustments probably by, by improving process rate and process times. These were the most important outputs. Thank you for your time. If you have any questions, please, this is my email to contact me.

[01:12:58] Roberto J. Arbulu: Thank you so much, Guillermo, and also Corey. I think it’s impressive work coming from two different universities in collaboration with PPI, you know, a young person like Corey being from the university, they are now working in the industry with all these ideas about capacity utilization.

[01:13:20] Roberto J. Arbulu: Guillermo, a bit more seasoned person professionally speaking in terms of, you know, having work in construction. It’s also learning this at UC Berkeley, and incorporating that into his personal toolkit. Yeah. If that makes sense. Right. And so from a PPI perspective, I just want to say that this is what the institute is looking to achieve with universities and younger generations.

[01:13:48] Roberto J. Arbulu: Right. And the concept associated with Operations Science and the methodologies like PPM and science. You know, are being introduced in our younger generations that are coming as engineers, as construction managers, as project managers to the industry. So, excellent work to both of you and keep it up.

[01:14:10] Roberto J. Arbulu: Keep up the work, and thank you so much for being with us today. We really appreciate it. Thank you. Thank you. Thank you both. Before we move into the next session I’m going to still, let’s say a couple of, I know we have Martin Fischer already with us, but I wanted to address a couple of comments I believe we saw in the Q&A here.

[01:14:43] Roberto J. Arbulu: The first one is a question. It’s, I believe, about the presentation McKenzie did. This is great. I’m not completely clear whether the impact slides were potential impacts from the simulation or whether they were implemented and achieved. I think the short answer is the impacts were really a combination of elements that were implemented, like they physically added more capacity in some cases, and were able to remove some capacity.

[01:15:12] Roberto J. Arbulu: Right. And in other cases, because you’re talking about looking forward, is what the detail twin was actually proposing as the optimal approach. And so it’s really a combination of both. The next question is about how long does it take on average to implement PPM, more specifically the five levers and the digital twin on a complex project?

[01:15:37] Roberto J. Arbulu: I can provide my opinion on this particular question and the digital twin can be implemented. I would say in about four to six weeks total, including an analysis done to this. It could be faster, but if you’re talking about a complex project, which is a question, it depends on how many production systems we are talking about.

[01:16:03] Roberto J. Arbulu: Right. But roughly, just to give you an order of magnitude, within four weeks, about a month or so, you could have literally a digital twin up and running with the technology that we have access to today. Okay. So I just wanted to, to address those two, those two questions.

[01:16:19] Gary Fischer, PE: Hey, Roberto, can I throw in something on that?

[01:16:22] Gary Fischer, PE: We found. Yes, please. Trying to use that or experimenting with that. In Chevron, what really drove the timeline was our ability to provide the data and the information that was needed to do the modeling. Right. Right. It wasn’t the modeling, the modeling itself is actually surprisingly simple to do. I, the experts can, can pull it together very quickly if you have the

[01:16:42] Roberto J. Arbulu: data.

[01:16:43] Roberto J. Arbulu: Yes, yes. And one thing that, by the way Guillermo and Corey did a part of their studies is, is, you know, the data gathering process and, and a bit of the translation that they, they had to do in some cases, right? Yeah. And particularly Guillermo, I want to speak for him on this one. Sort of he learned by doing, and he realized that the industry in many cases, you know, when you talk about capacity utilization, they don’t really understand what it means.

[01:17:10] Roberto J. Arbulu: And so you cannot come forward asking, “Hey, can you please give me, you know, what is the current capacity utilization rate?” That it’s, it’s not a, it’s not a, you have to answer these types of questions in a different way. Right? They might not be thinking about cycle time. So you have to ask about probably some rates, right?

[01:17:29] Roberto J. Arbulu: And, and the capacity number of crews. And, and so there’s a bit of translation sometimes that needs to be done. So I think that’s a great point, but it’s not that complicated. Right. And, and the value, it’s, it’s much larger than, potential time that needs to be spent on doing some translation on one piece of data to another piece of data.

[01:17:49] Roberto J. Arbulu: So Gary, I think over to you for the next session. Hey, very good. Excellent. Thank you. Bye-Bye.