Flex Your Digital Sim: Using Simulation to Optimize Your Operation and Workflows

JACOB NOLTNER: Hello, everyone. This is Jacob Noltner. This is our Safe Harbor Statement, and I'll let you guys read that. Moving on to our presentation here, the presentation of Flex Your Digital Sim, presented by myself, Jacob Noltner, and my partner here, Thad Perkins. We are both from Jacobs Engineering.

I'm a mechanical engineer. I have about 13 years of experience in the field and about six using FlexSim. And Flex Your Digital Sim is the title of our presentation today. And it is about FlexSim, a new software that Autodesk has acquired. My name is Jacob Noltner, and you will also be hearing from Thad Perkins, my partner here, and we both are from Jacobs Engineering.

I'm a mechanical engineer, and I've been in the field for about 13 years and been using FlexSim for about six. And I'm excited to share my experiences with you guys. Take it away, Thad.

THAD PERKINS: Hello, I'm Thad Perkins. I'm a manufacturing integration and optimization SME for Jacobs with a focus really on automation. I'm a mechanical engineer. I have a diverse background.

I've worked for a major consumer products company, worked for an OEM, and now I've been in consulting for the last 15 years. So looking forward to sharing information with you. Thank you.

JACOB NOLTNER: So on our agenda today, we're going to introduce you to FlexSim and what a digital twin is. And then we're going to move on to creating a model in FlexSim and how that model might interact with a control panel and process flow and some quick tutorials along with that. And then what do we get out of the model once it's finished?

Use some dashboards and what we can put into the model for a future state of your simulation and then some other tools that we can use to get better results or different results. And then we're going to talk about some industry successes that we've had since using FlexSim So what is FlexSim?

So 3D simulation software, they advertise themselves as discrete event. And what that means is that it's-- each individual event happens separately, and it's not just a spreadsheet of data. So one event can affect the next and the next and the next. Data driven, pretty much can use in really any industry. I haven't found one that doesn't work for yet, some better than others.

Leveraged to show, predict, optimize, flow and layout; easy to use; user-friendly; and customizable down to the semicolon, if that's what you need to do. If you're a very good coder, you can even get down to that level of detail if you want. But it is flexible, and it is simulation. So FlexSim.

Some applications we can use them in. Like I said, any industry, we've used it in the automotive and EV markets, consumer goods a lot, paper a lot, food, pharmaceutical, health care. I define pharmaceutical as like building the-- a vaccine or building a new medicine or that process flow of making an actual product.

Health care would differ in my eyes more of like a clinic and watching people walk through the clinic, that kind of thing. Restaurants kind of in that same vein as people moving rather than a product through a space. Airports, rail yards, and seaports are also very easy to be programmed into the software. Any process can be simulated-- warehouse, production, packaging, and people movement we've all done.

Any size, so we've done micro simulations down to one intersection in a plant to track traffic and how people and forklifts interact down to multi-facility campuses and how the different buildings on the campus interact and traffic in between them. Any level of detail, and we've done many different things with FlexSim. But some of the big ones that we use it for are identifying and eliminating bottlenecks, maximizing facility throughput, right size asset expenditures, minimize labor requirements, understand relationships between interdependent processes, and improve efficiency and safety, and minimize impacts of process upsets.

And then entire life cycle, digital twin, and we'll touch on that in a minute here. So how much detail do you put into your simulation? It's somewhere we always start and ask the client. What are you trying to get out of it? If you-- the more detail you're going to put in, the slower your simulation is going to run. So think about that when you're building your simulation.

Also, the faster it runs, the less it's going to cost. The less information you put in there, the less it's going to cost. Also, the more you put into the simulation, the more detail you're going to get out of it cost wise. So poor data quality can limit the benefits of a high detail model. So when you're getting your information, usually you're building this simulation from some starting point.

So you're going to have a number of data points that you can put into your model. And the quality of that data is paramount to get good results out of your simulation. Add detail to reach your desired confidence level. Obviously, the more detail you put into your model, the more confidence you're going to be-- confident you're going to be with how that thing works in real life.

You can also add some other things to the model or change things in the model to increase your confidence level, such as changing the numbers slightly, like 5% or 10% up or down. And if that totally crashes the model, well, you're probably not as confident as you would be if things didn't change quite so much. And you might think, well, obviously more detail is better, and that's not always the case.

It kind of really depends on the question being asked, and so we'll have two different questions here. The question, on average, can this workstation handle X products per day? You might only model in a product entering the workstation. It delays there, delays there, and then you move it on to the next thing. Very low code, very easy to do, and doesn't slow down the model all that much.

But if the question is, how do I double the productivity of each workstation, then you've got to detail or put in the process flow every step of the way, and you can try to figure out where your bottlenecks are in that workstation. So two different questions, two different ways to answer them, and two different right answers, really.

And always, garbage in equals garbage out, so always get good data. Otherwise, you're going to have a bad model that doesn't help anybody. So digital twins, it's a hot topic that people throw around sometimes, and maybe we don't all know what a digital twin is. But we like to think of FlexSim as a static digital twin or your starting point, your base as your digital twin.

And you're going to have highly detailed working environment of actual events. So it's probably something in the past, if you have real data, and then moving forward to a prediction of the future. Excuse me. As you build your model and you finish it, you might add things like sensors or other ways to read what's going on in your process.

And then from there, you can tell your process, is this good or bad? So giving it feedback as to how it's reading the data, and from there, you're really turning the FlexSim model into artificial intelligence. The machine is learning based on what inputs you've given it. And then you have a dynamic digital twin at the peak.

A different way to look at it here is you're going to communicate adjustments and make decisions based on what you've seen in the field. Compare those results, and then we get in a loop here of simulate live data, analyze the trends, and then circle, circle, circle as we get more and more efficient with our process.

So without further ado, let's start building a model. I think that's why everybody came and wants to see what we can do. So-- model space, this is where we build the physical model of your process. Several different file types we can use as background, including AutoCAD drawing file, or a PDF. Also, picture files work, easily stretched to scale to be the actual size of the model.

And we also have a wide range of objects preloaded so you don't have to put in an exact model of what you are using. FlexSim has those-- has basic models for you to use. And honestly, the basic models are better because they take up less space, and they are faster to use.

So I have a little video here of some example items that you can use in FlexSim. So this is a fork truck that we can use. You can-- if you're using a ship or a shipyard or rails, a lot of different items there, operators. They can serve a multitude of different functions.

Also, different machines, and it acts as a machine on the floor. They call it a processor. There's a multitude of different things that we can use, but those are just a couple quick ones that I'd like to show you guys. And so process flows, this is where your code is going to be put. It's used to build the logic of the model.

And pretty much every model that you do is going to have some kind of process flow. The only example I can think of that it wouldn't have a process flow is a very basic conveyor model as one thing goes on one conveyor onto the next conveyor into a machine, into another combiner, whatever. It's-- might not have a process flow, but if you get any kind of detail or extra, special stuff, it's going to have some kind of process flow.

Process flow is very customizable and easy to use. If you're not a programmer, don't be intimidated. They have all these different blocks that you can use one after the other, and it shows you in space how to-- what's going to be happening in your process flow. So it's very visually stimulating to see where it is in the process as well.

The FlexSim health care module uses an additional set of tools in the process flow. It focuses on people being the product rather than a manufactured product through a warehouse, if you will. And they kind bunch them up so that it's easy to use multiple items as one. And like I mentioned before, you can code right down to the semicolon and put a custom code in.

These are used very intermittently. It's maybe 95% of what you're putting down is not going to be custom code. And the custom code you do put down is likely a repeat of a custom code you've already done somewhere else. So I wouldn't get intimidated about having to use custom code, but it is there if you need to.

So we can set up a control panel. And we do this often with clients that want to be able to use the FlexSim model after the fact. This is a way for them to be able to input new inputs in without being afraid of touching the actual code of the model. So only changing stuff that they want, that they need to, and not breaking the model in the back end.

So we're going to do Using FlexSim and walk through a model. I've built this model already, and it's a pretty basic model, honestly, but it shows a lot of different ways that we can use FlexSim. I've made this. Felix Simpson is his name, and he's going to go up against a green machine here to start off. It's the old man against the locomotive kind of scenario.

So I'm going to hit go here and walk you through what he's doing. Is it going? So his job here is to pick up two different boxes-- a red one and a blue one-- bring it to his table, and then combine it into a purple one. The machine below is making a yellow box and a blue box into a bigger green box.

Once they have four completed boxes in their final stage, they'll be picked up by AGV and taken on to the next station. You can see the process flow there on the right. Usually, you see a token in that space, but it's a little green dot. Because I'm moving the camera right now, it doesn't show up. But it will once the camera stops, and I'll talk you through what he's doing.

But you can see my mouse there. He's dropping down the box and walking back to start the process over again. You see there's some green and yellow and red on the floor as well. That's called the heat map. And that's a nice visual we can use to see how operators move or where they're mostly moving.

So the more somebody walks on that spot, the more-- the hotter it's going to get or the more red it's going to get. So the camera stopped. You can see the different tokens there on the right. Each green dot is representative of a process trying to be done. Some of those tokens are stuck because we only have one Felix and we're trying to build more boxes than we can.

So you can see how a bottleneck is visually shown there. And then that token is representative of Felix. Now it was given back, and now he's traveling back to the red queue to load a red box. And then it'll click down, travel to the purple table, unload at the purple table, and then back to the blue box.

And he'll load that back at the purple table and travel back to the purple table and load it there. Then he's going to combine the two. And that's just a delay that we put in the model to represent him combining them. And we actually destroy the two boxes and create a new box, and that kind of is representative of the combination.

And then we change the box to be that purple color. And then that travels to the queue, and he unloads it. The conveyor is there on the bottom. You can see that they're building up four, and it releases four at once. That green line in the middle is a photo eye.

And so as the boxes cross the photo eye right there, yeah, it turned yellow. Once it-- if it turns yellow long enough, it'll turn orange, and that means it's been covered. And that is cue to turn on the conveyor in front to release all four of the boxes.

So it's uncovered now for long enough. It'll turn back to being green, and the rest of those four boxes will go down into the queue and call for an AGV on the back end. Felix is finishing up his fourth box, finally, and it'll call for an AGV on the next slides here as we move forward.

I believe that's all we wanted to show you here. The AGV picking up. So-- yep. Moving on, so we're going to have the same process flow here, but we're going to show you the next steps. And this is the AGVs filling the flow racking, and the different arrows are showing how the tokens are used in the process flow there.

Again, the tokens don't show up because the camera moves. But once the camera is done moving, you'll see the tables fill out on the top there and the tokens move through the process flow. So we're showing the green queue and how it's set up to build the tables out there on the right-- or, I'm sorry, the tables on the left.

I'm going to hit Run, and it skipped forward. So we don't have to watch the whole thing over. But we're picking up from pretty much where we left off on the last one where the three boxes are let go and then picked up here on this side.

So we're going to see-- my cursor is going to be the green token while the camera is moving. So you see the four boxes got to the queue, and the was called to come pick them up. They'll-- those four boxes would be on that table that I just circled. And then-- yep, so the AGV got there, picked them up. And then the AGV travels to the slot it found.

So it found a slot, its rack address, and then it says AGV travel there. Assign labels, and then it's moving the object onto the flow racking, and there's a loop there. We wouldn't necessarily need to put the loop in there, but visually, for this simulation, it seemed pertinent to show that.

But you can loop the loop four times and then the decides in there. If it has gone through four times, move on, go down, and then release. The release there is important because it's releasing the AGV back to the pool. So there's, I think, four AGVs there. And you acquire an AGV at the beginning, and then it releases back to the pool when it's done being used for the next thing to be able to use it.

You can see the fork trucks picking up the boxes here as they load the trucks into the next simulator-- the next video that we're going to show. But for now, we're just focusing on the AGVs. So get back up to the top here, and we can finally see our token and our tables. So the tables are already being filled out. They just didn't show up when the camera was being moved.

But you can see there's already tokens going through there on the right. And the tables are-- there's three boxes in the purple queue there, and there's three things on that table on the top. And those are kind of matching up. When those four green boxes reach its queue, the green queue, they'll show up on the green table there on the bottom as well.

So you can see them come on right now-- one, two, three, four. And then the AGV is called because there was four boxes, so there's the token there on the right. AGV travels to pick them up, pull, and then when it does that pull from the list, it's basically taking the boxes off of the list and finishing its travel down to the racking.

The pull from list is really an easy operation to use so that you can use the conveyors and the process flow in tandem. If you want to use one or the other, you could just do it all, but it's kind of nice to be able to use both. So the [INAUDIBLE] pull from list is quite essential sometimes.

So, as I described before, I did the loop in the boxes. AGV is released back to the pool, and we see an AGV1. It's got the purple boxes and doing the same operations that the first one did, just the start is a little bit different, obviously, because we're coming from a different place.

I'm going to move on. I think that's about all we wanted to see here. So the last part of this is filling the trucks from the racking that we just filled up. I'm not going to move the camera around as much. This is more about watching the process flow and how it fills out the table.

Another feature we can use here is zones. Zones can be used in really two ways. One, it can be, once you enter the zone on the process flow, you can limit the number of tokens that are in the zone. So you enter the zone, and there's an exit zone. So in between those two triangles, the yellow triangles there.

You can limit the number of tokens that are allowed to be in that space. You can also use the zones for a data collection point and use it into the dashboards we'll show at the end. But it can basically count the number of tokens or the frequency that there's tokens going through there, and we'll show that at the end. Also, we have variables here, basically, data that can be shared between tokens, which also can be very helpful in different situations.

So start this so I can fast forward it, which I just did there. And I pulled the-- yeah, so we're up to where we need to be now. Set it to times two speed. We've been watching it at one time speed, but this is easy enough to do at two times speed.

So we see the semi truck show up. We created the semi truck in the process flow on the right there. It's treated as an AGV. It shows up to the dock. It's a green token there.

So we're showing the variable value, that it's set to zero at the beginning of the process flow. Token is now down into the zone. And we're finding items on the rack, bringing the items over to the truck with the load, and then travel to the semi truck, move the items from the transporter onto the semi. Then we push to the list.

Basically, we're using the list as-- keeping track of the boxes or the products on the truck. And when it gets to a certain level, then we can decide to say, OK, the truck's full. Let it leave and bring in the next truck. You can see that these tokens are really looping inside this inner loop, if you will, and that there's a limit on how many tokens we're letting be in that zone.

Just one right now, because we don't want the fork trucks to be tripping over one another. So once the first token leaves, then the next one is allowed to go down through the zone, and now we're batching the tokens. So there's one all the way down. We're waiting for the second one to show up at that batch at the bottom.

And once the second one shows up, then it's allowed to continue moving forward. The AGV is allowed to exit, and then we start all over again, back at the top with the next truck. So that was quick, but I hope you guys enjoyed that, picking some of those up, some of that information.

So now we're going to work on what do we-- this was fun. We built the model. What do we do with the data that we can collect out of it? So these dashboards, so these are some very common ones that we use, utilization being probably the most used thing that we track.

Felix there on the left is 100% utilized, and that's basically all of the different colors added up to equal the 100%. The blue color doesn't count towards the total percentage on the rest of those. That's why you don't-- that's why that percentage is less. So it's the AGV1, AGV2, AGV3. They're only really less than 50% being used.

And some of those transporters are under 50% utilized as well. We don't really like to see anything over 80. That usually gets pretty hard to sustain in real life. Also, you can have the travel distance there. This was run over an hour, and we see that Felix walked 0.83 miles in that hour.

So in real life, that's not sustainable for a worker to be doing that much work or that much travel in an hour or throughout the day. The gravity flow rack graph there, you see minutes on the bottom and how many packages are on that rack and then some inputs and outputs. The inputs are the amount of boxes we were showing coming into the model, and then the output, how many left each of those different places-- the gravity flow rack, the different cues on the floor, and the final shipment number.

And that zone that we talked about in the process flow is shown there in that middle box. That's-- we were basically tracking the amount of time the semi truck was at the loading dock. And we can track the minimum amount of time it was there, the maximum amount of time, and the average. A lot of times that's valuable information that clients are after.

So how do we help Felix with his predicament here? Like 100% utilized, walking almost three miles in an hour all day long. So we're going to come up with some ideas that might help him out a bit as well as we want to help Felix out, but we also want to help the bottom line as well.

So we're going to do a future state analysis. And basically, I added three more Felixes. And because I added the three more Felixes, it was pertinent that I add another AGV because I saw a bottleneck there. And then I saw another bottleneck at the fork lifts, so I added another one of those.

And then we can analyze with the same tables what the added output was. So we have the purple operator increased from future to-- from current to future. You can see that highlighted there at the bottom. And so we can say that that's what we gained out of adding these three or these other things to the model.

So the problem that we're seeing here still is that the gravity flow rack is increasing and never coming down. It's going to eventually be full and not be able to continue past the hour. So we might need to level that out a little bit going into the future. But that's where our expertise in all this can help. So as you get into it, you'll see things like this that are obvious bottlenecks to you, and you can fix them as you move through.

Sensitivity analysis-- there's a certain combination of inputs that are-- or outputs that are unfavorable, and you can change your inputs by 5% or 10% in either direction. Or maybe there's a combination of percentages that you added. If it totally makes the simulation fall off the rails, then it's not really sustainable in real life.

But we try to do it for our clients to make sure that we're not setting them up to eventually fail. I mean, they might start out well, but if something goes wrong, they're not going to be able to get back on track. So we like to build in a little bit of buffer.

So the MMTR and the MTBF, both basic mean time to repair and mean time between failures-- is a study that we can do. And it's basically the scheduling in downs, and we can do it at a random amount of time. But you can make your machines go down for a certain amount of time at a certain frequency.

In this model, the transporter was down for a little bit that I have circled there. And you can see that the transporter, when it goes down, then the gravity flow rack peaked up because the one transporter couldn't keep up with the rest. And so it took four minutes for that gravity flow rack to fill up, but it took us all of a half an hour to get back to constant state in that green arrow there.

So that's also a bit of analysis that we can do and help out with as we're building these models. It'll tell you how long you have on a failure before you have to shut the machine down and how much it takes to work off that product.

So some project examples that we have. This one was-- I'm going to show you this basically for scale, and then we'll talk about some more projects that we've done. But this was an actual project that we did, and I scrubbed it for proprietary reasons. But basically, we can call it a facility that makes toy cars.

So you can see some of the stuff that we did as-- we see shipments come in here on the receiving end-- delivery, some boxes, some fork trucks coming to unload the truck. A lot of those packages were sent to the inspection lab before they were put off, but a certain percentage were just put right directly onto the warehousing shelves.

So that you can set up with your percentages or randomization as they let you do. Then this is like a chassis build area and engine. So there's more deliveries being shown here on the-- coming in on the south end of the building and then these different lines coming over. A gantry crane picking up these built motors and chassis onto carts.

The carts get pulled over to these assembly lines on the right by people. And then as they go down the assembly lines, they're onto this conveying system. And then all kind of combines together into this down spiral to the robot, which will move them from the down spiral over to the test cell.

It's called-- you just see the machine there, and it's basically representative of a person testing the final product before it's shipped. I think here you're going to see a work truck pick up or maybe take some stuff out to the shipping yard. Yeah, so we see fork truck four there moving to the right. It will go outside, drop off his load.

But yeah, this is just basically a full model that we built, and the customer is very happy with the results we were able to give them before they even built the building. And so a lot of the information we got was from a previous building they gave us. So it was a nice starting point we had as well. I'll let Thad talk for a little bit here.

THAD PERKINS: Yeah, I'm going to talk about some project examples, the success stories that we had. The first one we call zero turn, zero efficiency. This was a project for a client which manufactured zero turn mowers. FlexSim was used to validate the recommended manufacturing and operational changes prior to the implementation. And it also enabled an objective decision-making process to vet through the various options and to drive to the optimal solution.

There were three main project objectives associated with this-- throughput increase from 320 units a day to 620 units a day. No building additions or major building infrastructure changes were allowed as part of the scope and no changes in the operator headcount. So pretty tough criteria to work through. The project solutions, we reconfigured the whole facility layout.

We modified the assembly processes and procedures. We replaced or upgraded the required manufacturing assets. We were able to rightsize the equipment by optimizing the manufacturing lot sizes, and we used the optimizer module functionality within FlexSim. And the key thing about that is we were able to get 10% increase in productivity just by changing the lot size and not having to buy additional assets.

Material flows, we assessed and optimized and streamlined all the material flows. Material movements were automated so that to eliminate necessary-- unnecessary work by operators. New paint booths were incorporated. As we did the analysis, it showed that was going to become a bottleneck for the process. New assembly lines were laid out, and we incorporated a lot of new processes and procedures associated with subassembly.

The real focus was to move all the independent tasks that we could off the assembly line so that we were only doing dedicated, dependent tasks on the assembly line. And new operator parameters were established to maximize the efficiency of those. The results, we beat the target. We got-- 651 units were achieved. The objective was 620.

We saved the company. $350 million annual revenue was recognized, and one of the biggest compliments was from the client. At 60% of the construction completion, the client had already recognized an 80% increase in productivity. So a great project great example of how FlexSim-- the power it has.

The next project-- show me the money. That project was with the same client. Based on the positive results that we conducted on the zero turn business, they came back to us and wanted to evaluate their manufacturing process for the snowblowers. So similar to that study, we used FlexSim. We validated the manufacturing operation changes prior to implementing and, again, used it to vet through to get the optimal solution.

The objectives were different on this end. They were looking at optimizing the facility layout and manufacturing process through reconfiguration. They wanted to support a two-shift operation to produce 1,200 units a day at a 50% reduction in staffing levels. So the client was having difficulty in hiring resources and maintaining a consistent work staff for a three-shift operation.

So they wanted to get down to two but reduce that staffing count and maintain the current production of zero turn mower parts manufactured within the snowmobile-- snowblower facility. They did manufacture some of the parts they-- for the zero turns in there. So we had to maintain that.

So the solutions, we validated their capital plan. It was 70-- just under $74 million, and that was the investment in new assets; asset upgrades; automation opportunities for manufacturing, paint, and assembly; and material movement. It was all part of that capital plan. Storage areas for the metal and part storage within the facility were optimized and reconfigured.

The facility was reconfigured to a preferred west to east reconfiguration. So we looked at different flows, and the simulation said west to east was the optimal solution. We used the optimal lot size obtained in the module with the results we saw in the zero turn. We utilized it and leveraged and got some of the same benefits that we'd seen previously.

We optimized the manufacturing process by balancing the work centers, the flow through the work centers. Optimized the dock door locations to enhance the material movement and shipment. They have a lot of materials that they use a lot of metals. So we optimized that.

Optimized the maintenance and tool and die areas to reduce the footprint and improve flow within the actual facility itself. We consolidated and optimized the quality inspection and test areas to get those more central to where they needed to be. And then we established new material handling methods to reduce associated injuries as a result of inappropriate or activities or repetitive activities.

And then we segregated the fork lift and pedestrian walkways to improve plant safety. They had a lot of incidences where there was interference between the fork trucks and the pedestrians. Final results on that, we validated the future state plant and manufacturing and assembly processes. We achieved the 1,200 units a day.

We were able to reduce the labor down to 42% of the goal. 50 was the target, but they were still pleased with the result. 50% was very aggressive, and we had an annual savings of about 13.5 million was recognized on that project.

So the next project we call AGV pool party. That's a project where areas of the customer's facility production field didn't have enough sufficient capacity to support their increase they were seeing in their business for both-- some of their current activity but certainly for the future product demand. So the effort primarily focused on the production of unit loads coming off the converting lines and packing lines and the movement of those loads by laser-guided vehicles into the warehouse.

So the project objectives were to identify and eliminate process bottlenecks in order to efficiently and optimize handling the increased loads for demand. And then they enabled the client to use the simulation outputs to aid in the justification of new and reconfigured assets to support the increased demand. It was used as their business justification.

The solutions, based on the process assessment and analysis that we performed by the simulation modeling, the key solutions and recommendations that came out of it, we upgraded the robotic palletizing cell or packaging cell to be capable to put in pallet insertion. We saw that as a bottleneck. We increased-- installed three new packing lines cells in between the robotic cells, and that implemented direct conveyance from the robotic cells to the packaging equipment cells.

We expanded their current fleet of LGVs by 4 units from 16 to 20, but that was a reduction from four of what the client was anticipating initially. They were expecting to have to increase by eight. We relocated one of the stretch wrappers at the end of the converting line to put it in a more favorable flow position. We eliminated the LGV barge system.

They were using a system that basically had a tug behind the LGVs, and we eliminated that and we eliminated all the packaging cells within the warehouse. So the results, we got 30% equipment savings over the projected capital plan. And we had a reduction of 50% of the new automated equipment required to be purchased that they were anticipating to have to purchase.

And the last project I'm going to talk about, we call BAD-trey use of space. This case study was for an expansion of a client's vehicle battery purchasing-- processing facility. FlexSim was used to validate the building size requirements to improve the process flows and safety and optimize the equipment in the process layout. The objectives were threefold-- to increase the facility throughput by 1,300%-- so a significant increase there.

Incorporate building expansion as required. Because of the business, they wanted to have the flexibility. And renovate the existing building as required. So the solutions, we validated the initial client layout and concept and some assumptions. They had done some work, and we were able to validate that. We standardized the approach for the battery receiving and prep to meet the new rates.

We utilized models to pinpoint high impact areas to maximize the use of the facility space. And we've confirmed the key project design criteria. The results, we proved that the current processes were not scalable to meet their throughput goals. We validated the existing building was insufficient to meet the client's goals.

We developed an implementation plan abandoning their prior ad hoc approach, much more scientific approach that we used. Excuse me. We improved the process ergonomics and safety. We rightsized the new building addition. It was a 112,000 square foot addition.

We determined the existing building renovations-- and that was about 6,800 square foot or 68,000 square foot-- and we optimized the floor plans, equipment selection, and headcount requirements. The bottom line on this project is they could not have sized the building, laid out the equipment, and implemented the necessary changes required by the project schedule without leveraging the simulation. Excuse me.

JACOB NOLTNER: Back to me. This project-- an apple a day, keeps extra staff away. Pretty different than most of the simulations that we've done. This was-- and I alluded to it earlier-- done in FlexSim health care module, and it's based around staff being the main product rather than a package moving through a facility.

It was a clinic that was very overstaffed and didn't use their staff in a beneficial way, such as a doctor would be taking the patients from the waiting room to their different operator or different clinic rooms. So you kind of want to match your staff with their purposes. And they were not doing that, and they wanted to level that in a better way.

And we did the simulation of their facility, and we could just basically, based on redoing their staffing schedule without changing anything else, we got a pretty significant staff reduction and a better use of space-- 48% and 35% there, respectively. And then based on the staff reduction, they were able-- and the change in staff that they staffed-- so probably more technicians and nurses rather than a significant number of doctors, that-- leveling out their staffing where they're able to save an annual amount of $400,000 there annually.

So that's a neat, cool, different kind of project that we were able to take on. AGV Avenger project, this is a client of ours that works in consumer paper products. And they're going from a staff of fork truck drivers from each machine back and forth to the shipping warehouse. And they were changing that over to AGVs.

They were concerned that the AGVs, that there'd be more of them because they don't travel as fast as the fork trucks. But because there'd be more of them, a number of the intersections in the facility would be swamped between the AGVs and staff at certain shift change times of day.

So we were doing a safety analysis of-- and this is where I got into we simulated just one intersection, all the data coming into that one intersection. We did the facility as a whole, but on some of the intersections that we deemed more hazardous, we dug into a deeper simulation and just simulated just the intersection. So that was a lot of-- we are able to count 38,000 risk adverse actions avoided annually within the simulation.

Storage capacity increases with the change 82%. And then the miles driven by their AGVs versus the fork trucks that they were using originally, 3,321 miles less driven. And we calculated that with the model as well.

Why so serious-- same client. They had a line that was being installed in multiple facilities across the US, and it was mostly conveyed line. And so it-- like a packaging line. And so the timing on all the different SKUs with the different photo eyes and how everything would start up and if there was a down, like how does that pattern change for each different SKU?

So we were able to achieve a 50% faster start up for each of the number of machines that were delivered to these different facilities around the US. So they were very happy with that.

THAD PERKINS: Yeah, I previously had-- this is one of the first case studies that I talked about. And this is just some detailed, more detailed slides. So as I mentioned, the throughput was a target to go from 320 to 360, no building additions and no head count changes. These were the solutions that I talked about already.

And here, you get and actually see what the layout looked like. So in this case, we went from an east to west flow. So the areas that you see in the green, those are machining centers. And in the yellow are more manufacturing process, welding booths, and different functions. The two purple things are the actual-- those are the powder coating paint stations.

And then you see the various assembly lines that you-- there were actually six lines that we had, and then the subassembly lines are outside of those. So that's the actual layout. Slide, please.

So the outcome-- 651 units, $350 million annually. I talked about the quote, about the 60% This is a direct quote from the client, and it does mention the client's name. This is Ariens. They do provide us the opportunity to use them as a case study and publicly say their name.

They also have been a very good client for Jacobs in the fact that they bring a lot of work in based on the performance they had with us. So we've executed a lot of work for other clients because of the Ariens work that we successfully executed. Slide, please.

And here is the snow blower. So again, these were the objectives. These were the challenges. As I mentioned, it was really looking at trying to reduce staff, not because of the fact that they felt they had too high a headcount. The issue was they could not get enough employees to stay.