In the third installment of the Stream / Podcast, we go over the following manufacturing technologies: IoT, Predictive Maintenance, AR/VR, and Big Data.
— Vladimir Romanov: I’ll do it so welcome everyone. This is episode three of the manufacturing hub in which myself and Dave go over some of the topics in manufacturing we’ve discussed up to this point, our experience. In the last episode, we talked about the careers, the job market, how to ultimately upskill yourself for manufacturing. And today we’ve got some for a very big, and I think ambitious topics that we wanted to discuss, which are going to be IOT, we’ve got preventive maintenance, and then we’ve got AR VR and last but not least, big data. So as you can see, Dave’s already wearing his headset. And, I don’t know if you want to make that the first topic or go into IOT.
— Dave Griffith: I was, I was actually going to say, can we make it the first topic so, so for anyone who is listening live, I am currently wearing an, each empty one by reel where, I’m wearing it on my head. You, there are a bunch of different configurations that you can have including on the hardhat. but for anyone that hasn’t worn them, they’re actually kind of heavy and not something that I necessarily want to, want to wear for the, the whole stream. and so if we can start with augmented reality and virtual reality, I think that would be a great place to kick it off.
— Vladimir Romanov: Well, I certainly think that you have a lot more experience in that domain based on the headset that you’re wearing. I, you know, I’ve personally only experienced it a couple of times at the automation fair and not even, you know, the one that just passed, but the one in 2019 where a Rockwell automation presented, you know, their, I believe the name is like who Foria chalk. I don’t know if that’s the right. Yeah. And, I think they had released, like a trial version or a reduced price cost version for the entire 2020, which, I had kind of applied for, but I never actually got before your studio in the end. So I didn’t have that much of an opportunity to practice. I think it’s still, I guess again, only just personal opinion. I think it’s still fairly early on to kind of see applications in augmented reality, but I think there’s going to be quite a few benefits as that technology matures, but I, I, I still think that it will be driven primarily by like the gaming industry, you know, to make that a lot better. But, you know, I’m very curious to hear your thoughts on it, Dave.
— Dave Griffith: Yeah. So I’ve been following augmenting reality, virtual reality in the industrial setting fairly closely, at least for the last four or five years. And for a while, I was very much in the mindset that you are a flat of. I think we’re some number of years away. I think what PTC has done with the suite, the chalk and the expert capture in particular are very interesting. And between that, and kind of everything that’s gone on with the pandemic, we’re certainly pushing what we thought, what we thought was going to happen, much sooner. and so, a little bit of a couple of, I’ve done a couple of projects, in AR and VR, and I’ve done a couple or at least know a couple of other companies that are doing some very interesting, that are doing some very interesting things. And so, I guess, I guess it was about three years ago, I did a project, inclusive of what we’re basically calling mixed reality, right
— Dave Griffith: It’s taking augmented reality and virtual reality kind of slamming them together. and we kind of built all of that on an ignition perspective. the, on an ignition perspective platform that we then use to help kind of push the, the rest of the project along. And it was, it was a landfill project and one of our partners at the time actually took a drone. They do a lot of drone services, so they flew the entire, landfill, and then they went out and they uploaded it all to a Pix 40, which basically rendered it into this complete 3d model, which as we are going to help push the pilot project to the next 28, facilities in this County, we, they then literally took like a hollow lens and allowed everyone in the boardroom to be able to walk around the facility. And, there was like projections and pushes to be able to like basically take SCADA information.
— Dave Griffith: So like temperature of the flare station and pressures and other things like that, and actually overlay those into the 3d environment to give you the mixed reality. and so that was a very interesting project. I think that, that, w we want a firebrand or, ICC, I guess a couple of years ago on that. And so it was an interesting project on it worked well on the marketing sales side. And I think that we could certainly see more of that moving forward. And I think that it’s still very, very early, in the industry as to, as to projects like that, or people willing to commit the money to, to actually do that for all of their facilities. And then go ahead.
— Vladimir Romanov: I was going to ask just so I understand like the use case. Well, so what you’re saying is that they permanently, I guess, may the facility in the virtual environment, and then you could monitor the facility like from, I guess, anywhere, but that was like an ongoing thing, or they wanted to create something for, you know, structuring the, the site or what’s the, like, I guess what’s the use case exactly. In terms of like long-term usage. So
— Dave Griffith: The, the use case was more of the, we have done a bunch of these upgrades, inclusive of some SCADA, inclusive of some ignition perspective. How can we show what we’ve done to the people that sit in a boardroom And so that was, it was kind of an add on service of the, the
— Dave Griffith: Partners did a lot of drone services because they did a lot of drone services. They took the drone out, they flew it, and then they use some other services to render out this virtual reality environment. and then kind of the, the Lang over of the SCADA, it was similar to like the, the concept of geo-tagging, what was very much like just a test to show them that we could do it. I was not involved in the rest of the project. I am not sure. And I honestly wouldn’t think that it was a bot for, for the rest of the project. It was more of a, it was more of a use case to, to show people what the entire project look like and was inclusive of
— Vladimir Romanov: What, I guess, what are your thoughts on the, maybe a traditional production floor level like application Cause I know, and again, this is purely from what I’ve seen thus far, but there’s a lot of push in the, I guess the training arena rights, you should be able to perform certain, let’s say like maintenance tasks, whether it is, I don’t know, replace a motor or even, you know, look at a certain pulley, for example, it can kind of like show you here are the components, here’s what potentially breaks, you know, show you, a lot more information. Like, do you see any of that coming into play, you know, anytime soon. And if so, like how successful do you think in terms of like ROI is maybe that may be, if you can even kind of talk about that
— Dave Griffith: So, so I, I’m doing a project with, with one facility taking actually potentially some of these headsets onto the plant floor for their electricians and instruments staff, so that they can have all of their maintenance manuals, they can have all of their events and cause logs that they can have. I mean, inclusive of like your entire Microsoft environment on a headset like this, so that if you have a problem, instead of saying, okay, I have a problem with this motor. and then going to a computer that is probably a static computer or going to the physical books to look for it, instead of all that back and forth, you have the ability to call it up with your voice and go through and look at that you can pull out and look at schematics. it’s actually extremely, interesting. you can build workflows on these now.
— Dave Griffith: And so I think all of that is valuable. I see kind of a shift between going for headsets like this and going more towards that traditional iPad. I think that a lot of the older, subject matter experts, I think a lot of the people who aren’t comfortable, you know, calling out exactly what you want a headset to do would be more comfortable with the traditional tablet on the floor. So I certainly see a push towards that. there is a company, I actually know them fairly well based on outside of Atlanta called IQ agent and the owner, Bob Meads. He was actually the first person that introduced me to AR VR. He put a hollow lens on my head. Five years ago, we went and tapped around. We all look like idiots. We were at a bar, so it was, it was appropriate, at an event.
— Dave Griffith: And so I’ve been following those guys and those guys are very much the QR codes and the, you know, everything has a QR code. You can scan the QR code, you can pop up the instrument and other information that you have and need, from those. And I think we’re certainly seeing a lot more of that, catching on. And I can see that being extremely valuable, to go troubleshoot or be able to blow out schematics of very specific items. and then kind of talking ROI. A lot of the ROI knowledge that I have is based off of kind of what I see the PTC view for you, people talking about. and when they go through a demo, they’ve got this very large Excel spreadsheet that they talk about, you know, man hours and cost per man hours and cost for downtime.
— Dave Griffith: I think that th there’s a very quick and very large return on investment and because of the quick and large return on investment, I think we’re going to see more and more people pushed towards using things like that. And especially the, of euphoria chalk. And then before he expert capture, I think that as we lose subject matter experts on the plant floor, we’re going to need to find some way to empower the people who are actually on the plant floor or as we potentially roll into more lockdowns. every country is different. I know Canada, a lot of, lots of the UK are very much locked down. It’s it becomes the less of the question of, okay, how much would it cost to get this person to get on an airplane and fly here or drive over here. And very, it could very much be a can they physically do it Can they legally do it And so at that point it becomes the, okay, well, what if I can get someone to hold up their iPhone and basically FaceTime back and forth and the other person can be there and they can circle the thing and they can walk them through exactly what they need to do. And that becomes a th that becomes a very inexpensive process to go through and to get a suite like that set up and to get, you know, some $400 iPad set up.
— Vladimir Romanov: Yeah. And I think that’s, you know, what you mentioned is kind of the most basic form of augmented reality, right It could be as simple as like setting up a phone call with, just a video chat. And I think that would be extremely helpful. I can, for example, a test to that when I’m troubleshooting machinery and I’m talking to a technician, you know, three States across, and, he’s trying to explain to me what’s going on with this. Palletizer why the robot is not getting the right amount of boxes, but without actually seeing it and being able to like direct them on what to look at, it’s it becomes very difficult, but, no, I mean, like all good points and we got, we got a comment from Frank actually that I think, like I seen this application too. So he says that I did one of these with Raytheon main purpose was to be able to examine the machine by, around it, looking at mechanical assemblies to determine cable routing and et cetera.
— Vladimir Romanov: So, let me see. There’s like a Seymour. yeah, so I I’ve actually seen a similar, virtual reality room at, like P and G had like a site that had that, and they’ve, they’ve invested a lot of money into it, but ultimately what people could do is come in and for example, you could describe what you wanted a certain piece of machinery to do. And then the mechanical designers could like, you know, plan something out and they could present. So you could walk around the machine, so to speak and someone would control the machine, like on how it was oriented and kind of like, you know, take components out of the machine, counsels, certain, certain things out. But, I guess, I think it’s like, again, my perspective that on the plant floor, I think there’s still a lot of hurdles, right.
— Vladimir Romanov: And we haven’t even dove into the topics of, let’s say like privacy, because ultimately if I’m, you know, if I’m wearing this headset that can film everything, someone can check clinically, you know, get that data because it is going through the plant network and be able to like, watch everything that I do. And then you get on the legal issues, which again, I’m certainly not a lawyer, but I could see how, if I, you know, look at certain things or, you know, spend too much time, I don’t know, looking at something different than someone goes in and analyzes what I was doing every second, at the job, I could see how that could be a problem for, you know, privacy issues. And of course, again, safety as well, like if you’re wearing this headset and I’m sure that there’s going to be plenty of new applications, that people are going to find at the plant floor, you know, with these, with these things. And, you don’t want to be watching something while you’re walking across the plant floor. You know, that’s the, the, the biggest lawsuit that could ever happen probably to a facility. But, I mean, I dunno, I I’m, I’m really curious what, you know, what some of these big players are going to come up with, but I’m still thinking, you know, three to five years to see something really concrete on the manufacturing floor that delivers like an ROI that’s, you know, concrete and those applications and everything works as expected.
— Dave Griffith: No, no, I would agree with that. And I would say that when I look at applications like this and kind of all applications that I look at building, it it’s very much how can we enable the people on the plant floor How can we enable the people at the facility to be able to do their jobs with less physical stress on them easier. How can we get the information to them when they need it so that they can make better decisions And so it’s very much along the lines of how can we enable them. And, you know, there will certainly be use cases where people misuse, equipment, but people are misusing equipment on plant floors. Now, you know, you run into, you run into cybersecurity issues, of people, you know, plugging in an ethernet cable and dual homing a computer because they want to get on the internet from an operator station or plugging in a PS three, because they want to play Madden on third shift on Saturday night. I know people, I know people that have walked into that or worse. And so it it’s one of those that people are going to misuse and find ways to misuse as edge cases. A lot of this for me is how can we help take and push some of this on the plant floor to allow people who are going and want to use this to be able to do, to do their jobs in a easier, less stressful way
— Vladimir Romanov: No, for sure. And I mean, like my only comment is that most facilities that, you know, I’ve been to have banned the use of, let’s say cell phones outside of, you know, whatever the cafeterias or just the general areas, because of, it’s just too difficult to kind of monitor and make sure that people are attentive. So having a requirement of a headset that can potentially display all those same things is a little bit intuitive, I guess, from a safety standpoint. But as you said, I think there’s going to be a way, right. To like restrict certain things to kind of minimize access and what have you. So I think with proper measures, it will get there, but, let’s, let’s get onto, an even, I guess from my perspective, interesting topic and slightly controversial, IOT, internet of things, and maybe even the industrial internet of things as we’d call it in manufacturing. But, to start off, I guess, I’m assuming most of you are familiar with it, but, you know, just, maybe get your take on it, Dave. Like, what do you, kind of understand when someone mentioned IOT, but at the same time, even more importantly, what do you, make of it when it comes to digital transformation And then what kind of benefits do you see IOT bringing to manufacturing And I guess this has been going on for a while now, but
— Dave Griffith: Yeah, so, that’s good. And I would agree with that. I think most people should understand IOT. I mean, w when you can go to the store or when you could go to stores and you could buy, refrigerators that were connected to the internet, that’s just kind of a bizarre thing, but basically anything that is now connected to the internet, or can be connected to the internet, inclusive of smart sensors, and otherwise is your IOT or your IOT, your industrial internet of things. And so that is, that is kind of how I consider it. And I would say, you know, the industrial internet of things has been going strong for the last, you know, five plus years, or maybe closer to 10 since we branded it as IOT and kind of even before then, it’s very much the, what sensors can we put on the plant for and what wireless sensors are we using so that we don’t have to run wires to it
— Dave Griffith: I would say, especially in the last probably three or four years, we’ve seen a lot of companies come in and install smart with their data collection devices, which as we were joking about before we went live is basically just a hardened raspberry PI that we’ve connected some sensors to that we’re pushing up to the internet and it’s some sort of cloud solution. And then we’re running some sort of diagnostics across it that we may or may not say is, you know, machine learning or artificial intelligence. And it may or may not actually be machine learning or artificial intelligence. you’ve got your little Dell and you see, Oh, gateway gateway
— Vladimir Romanov: Units, right. Edge computing. But no, I’m sorry. Go ahead. Go ahead.
— Dave Griffith: No, no, no. And so I was saying it’s one of those things that has certainly been with us for probably longer than either of us have been in the industry. You know, there are probably sensors and other things in my parents’ basement that are 20 plus years old. And so it’s a, it, it’s one of those that, IOT has been with us for a long time, and it’s very much kind of the backbone of industrial automation of manufacturing. It’s how we collect the data to be able to make decisions as a, as we’ll talk about in a little bit, but what’s your take on it flood
— Vladimir Romanov: Yeah. I mean, I think that, in the general sense of that term, it’s been, like I said, a little bit controversial for me because I think that at first it started off as a fairly good like marketing initiative. And I think there’s a lot of benefit to, you know, making something connected obviously for data collection. But I think it somewhat with, again, in certain areas with certain vendors turn into more of a gimmick where everything has become IOT and that like encompasses any device that has like any network capability whatsoever. And I think it’s very important to kind of figure out the use cases on like a facility to facility basis, you know, when selecting some of these, these devices. And I think like one example that I will kind of maybe mention at the same time, it’s like IO link. And, you know, like that protocol I think is extremely useful, but it has certain cases in which, how to say it, if you don’t really leverage that protocol the way it’s meant to be leveraged, then it doesn’t really matter whether you have it or not.
— Vladimir Romanov: Right. So you’re paying for a feature that’s technically IOT capable, but if you’re not utilizing that to save cost on, let’s say like maintenance of, you know, like replacing those sensors when they break down, then you’re not really, getting any benefits out of IOT and similarly, even with, data and we’ll get to big data again, like that’s the same kind of a discussion, but, there’s a lot of, I feel how to say it, craze about getting the most data possible and trying to just shove everything into your database, make sure that you can collect it so that at some point you could process it. But the reality I feel is that we’re collecting a lot more data than we can process and then drive decisions from. And so that’s where I think like IOT has become like over marketed and over leveraged.
— Vladimir Romanov: Right. So it depends on the facility depends on the use case. And I think ultimately that’s why, like, in my original question, I’m like, I’m really curious to hear what you have seen in the industry where like there has been an application where we’ve gone does data and it really helped us solve problem, like ABC versus like we’ve got all these really cool devices and trust me, I love technology. So, you know, I could spend a lot of money on new sensors and stuff, but ultimately for a plant floor, it comes down to ROI. Right. So what are your thoughts on applications
— Dave Griffith: So I think that that’s a really good idea and I appreciate your calling out, IO link. So there’s a company that I’ve done a fair amount of work with back when I was an application engineer, we were offering banner engineering and so they do a lot of sensors. I do, you know, banner engineering as well, flat I’m, I’m seeing you smile. Okay. So, so they have a bunch of very interesting, like wireless and solutions, inclusive of some predictive, solutions. And so they, they do temperature, vibration humidity sensing, at a minimum. And then you can bring this back into a smart gateway, which you can then connect to the internet. And I believe they’ve got cellular connectivity to some of those that you can push up into a cloud-based solution. Like it kinda hits all of those IOT. it kind of hits all of those IOT buzzwords. it hits the cloud, it hits the predictive. And so a lot of that is based upon, you know, just your normal science of we’re gonna go, we’re going to measure some vibration of bearings typically on motors or blowers. And then based upon what that looks like, we’re going to be able to track the life of some of these and at least ideally be able to change out, be able to change out and do preventative maintenance or predictive maintenance.
— Vladimir Romanov: You seen that application actually like, sorry to interrupt, I guess your, your train of thought, but I’m curious if anyone has seen the application of vibration measurements. Cause I, I hear a lot about that and I’ve asked a lot of my colleagues, you know, no one has seen it applied at an actual facility and maybe it’s just an industry difference, you know, where food and beverage doesn’t use, the largest of motor. So maybe the cost is just not there. But do you have any examples where like, that was actually a big initiative
— Dave Griffith: Yeah. So I guess a couple of things, one, I know a bunch of banner engineering folks, and I know that they’ve done it successfully. I believe they’ve got case studies out. I I’ve had offers that to have some of them sent to me. I just haven’t had the right test case to be able to watch to test some of those apps, but I have seen an exceptionally large, solution. It was, so I wasn’t part of this project. It was a conversation that I had as part of this project. And so Emerson offers large vibration, predictive, solutions, for at least, at least for the paper industry. And so I was having a conversation with a mill and I believe they had like 15 to 20 lines and on one of their lines, they had like 600 vibration sensors. and they were looking to scale that from one line to all of the lines because they’ve run the calculations and the cost of them putting all of this in paying for what I assume is a yearly subscription and then predict, and then doing preventative maintenance on this significantly outweighs the downtime cost of some of these actually going down.
— Dave Griffith: I would say that when you’re doing vibration sensing, you know, it has to be on very high dollar motors and it has to be on something that we’re going to repair as opposed to replace. Like we’re not going to put, or we’re not going to put a up like this on, you know, a $200 motor, but if it’s a $30,000 motor and it’s going to take four days to get a new one, whereas we have a spare bearing, you know, we have a spare bearing that we can swap out or we’ve got a spare, whatever that we can swap out and do our predictive and preventative maintenance, then the cost of taking it down for an hour to do that, as opposed to, you know, we’re $50,000 an hour in downtime for the entire line. And now we’re behind on everything. It becomes, it becomes a very simple return on investment equate equation when you get to a lot of these high dollar value offerings. And at least my conversations that I’ve had with the banner people was very much along the lines of, you know, you have to understand, you know, who you’re offering this into and you’re not going to put, some of these vibration centers on the least expensive items. You have to be much more, much more specific as to, as to where you’re feeding into.
— Vladimir Romanov: Yeah, that makes sense. I think, while you were explaining, we had, two more comments that make sense and a tie into this conversation. So that mentioned that, he knows a rep from IFM and I think IFM offers some of those solutions as well. And then we’ve got, an instrument that’s set in the steel mill industry. They always have vibration monitors. So I guess like my takeaway from this is that it needs to be big enough, like you said, it needs to be expensive enough. And ultimately I think it’s also very prevalent in the industries where you really can’t stop production, right Like in, in food and beverage. And I guess in pharmaceuticals and many others, you can just kind of stop the line and kind of take a break from maintenance. And that happens fairly regularly versus in like a paper industry. And again, I’ve only been to a couple of those plants, but steel, a steel mill, I assume, is the same. You can’t really stop production. Right. You’re just pumping out the product and stopping would result in a huge loss and ultimately impact on, you know, on your production. So yeah, I guess I could see how that would apply. And what about, some other interesting IOT applications or, you know, cases I, I think you were on that train of thought, but I interrupted you with the motor question.
— Dave Griffith: No, you’re good. You’re good. And again, I think it’s who runs three shifts who runs continuous. And sometimes if you’ve got, you know, a couple of weeks or a couple of one week segments to do preventative maintenance, then it becomes less important. I was at a facility that makes aluminum cans for beer. Those guys have like one and a half down days a year. and yeah, yeah, I know it was, it was, it was a little crazy. but no, and so I think that we’re, we’re certainly going to see more of those solutions. I certainly see a lot of like IOT offerings. That is almost a, and I know we talked about MES manufacturing, execution systems, and one of the earlier streams, I almost see a lot of offerings like that, that are trying to pull the process information away from the plant floor.
— Dave Griffith: And I am certainly a little bit tentative about IOT solutions that are like that or IOT solutions that promise that, you know, they’re doing machine learning and artificial intelligence. It’s, it’s like, anytime I talk to, I’m gonna make a joke about sales guys. Cause I am a sales guy. And so I feel like, again, it’s like anytime a sales guy says, Oh, we can do this for everything, every application under the sun, you’re like, I’m going to have to dig into this a little bit more and see like, what are you actually doing Where are you guys And actually I do you have the, you know, experience needed in order to actually get there Or is it just like marketing and sales jargon So I feel, especially when you get to like the IOT or the IOT space, if it’s not a long-term company, you have to be extra weary and do your due diligence before jumping into a solution that promises to fix everything.
— Vladimir Romanov: Yeah. And I think I’m also on the cost side, you know, as we spoke before the stream, I think, one takeaway from this as you gotta be really careful with what you ultimately buy, because there’s a decreased, I guess, price on the hardware that is connected. And, you know, speaking of the raspberry PI that we mentioned at the beginning of the segment, they’ve released their new version for $4 U S right. I believe it’s like one week or two weeks ago. And, like the cost is just continuously dropping and you can do amazing things with just that one raspberry PI versus, you know, some of the vendors and, maybe some smaller companies will try and sell you solutions that, aren’t necessarily cost-effective, but are also, I would say like the next caution is, closed architecture, meaning that you can’t necessarily use them on your plant floor with a wide range of equipment, right
— Vladimir Romanov: So you’re locked in purely into their ecosystem. You can only do very basic features that only work with their hardware and software. And I think that, that becomes extremely frustrating for those who may not be aware that that’s the case, because as you said, they might be sold on that idea of like IOT conductivity. Here’s your beautiful gateway that can send your tags from your floor devices to, to the network. And then only to find out it’s like extremely limited locked out and you can’t do anything without calling them and getting them on the plant floor where, you know, you’re getting billed hundreds, if not thousands of dollars to, to get support calls. But yeah, it’s a, I think it’s a interesting industry we’re for sure. Going to see more and more changes, and I’m curious to see what’s gonna come next.
— Dave Griffith: No, great. I think that we’ll certainly see a lot of people continuing to push into it. I am excited to see all of the startups, especially getting into the hardware space. and I think a lot of them are doing some very interesting work. And I think that we’ll certainly see more and more opportunities in the IOT space, software hardware kind of along the lines in the, in the coming years, which is, which is certainly good, you know,
— Vladimir Romanov: And we’ve got an interesting common. So someone says the company I work for now makes a lot of IO link devices and IO blocks. They’ll talk to both PLC for primary PLC IO plus integrates, cloud communications using MQTT protocol and some of the other platforms, which like, again, I think all those teachers are great. but I think your manufacturing facility needs to be at a certain stage where they can leverage that. Right Like even if you’re an MQTT broker, for example, with your sensors, that doesn’t mean that the factory is going to be immediately ready to receive that data and even process that data into something meaningful. And I think, you know, back to both of our points, you can be sold them that solution simply because it has all these quote unquote IOT features, but you may not always be able to leverage them immediately. And although they might be really good for someone that can, it’s not always cost for like a specific application. So that’s, I guess we can debate this for a really long time, but that’s where I’m going to leave it for, for now. But yeah, really curious to see what’s going to happen there. Okay. Predictive maintenance, we’ve touched a little bit on it. but do you want to, start us off with some kind of a definition and maybe again, some use cases of, what you’ve seen and where’s that going
— Dave Griffith: Yeah, yeah, no. So I like that, I guess, I guess the best way to talk about predictive maintenance is what happens before for predictive maintenance, right So like at the most baseline we have the it’s broken. We have to fix it right now, mate. It’s right. The, the it’s on fire and we need to fix, and then beyond that, we’ve got preventative maintenance, which means we’re doing things at a variety of cycles of, you know, Hey, we know that we need to change this out, you know, once a year. So let’s say you’ve got a wireless solution. It’s in a remote site, it’s a battery. You know, the battery should be good for 18 months, but you change it once a year just to just be careful. predictive maintenance is kind of that next step of work pulling in collecting data. And we’re monitoring, especially a lot of these high dollar value items so that we can go ahead and do our preventative maintenance before we get to the point of it breaking. And then the facility or the line goes down. And in many cases, if something were to break, maybe we have to change a bunch of other items. And it certainly takes longer to fix the broken thing than it does to fix the less than broken thing.
— Dave Griffith: But, but no. And so I guess after I, after I kind of define a little bit about predictive maintenance, I should say I’m super bullish on predictive maintenance. I, I very much think that it’s the way that we’re going to go. despite the fact that probably at least twice a month, I talked blue in my face about how we should track our maintenance and how we need to do preventative maintenance. I feel very bullish about predictive maintenance, and I feel that as we’re taking technologies onto the plant floor, and as we’re going through the process of expanding many of these offerings, we’re going to be able to integrate predictive maintenance, predictive maintenance programs as we’re going through digital transformation and industry Ford auto initiatives. We’re certainly seeing, or we should certainly be talking about predictive maintenance
— Vladimir Romanov: Specifically, I guess, like I want to get into the weeds of this conversation, right Like, let’s take a, some kind of a packaging line, for example, like what would you see or what would you need as let’s say the solutions architect to give them a better view of their equipment so that they could perform, let’s say better maintenance and then cut on some of the costs. You know, like I’m curious about the specifics, I guess, of what your vision would be.
— Dave Griffith: So I guess in my like ideal vision, it’s, it’s a, you’re actually having a conversation with the machine builders. And like, let’s say we’re building a facility from scratch so we can put new equipment in, which is obviously going to be the most ideal case. And then so hopefully these machine builders should have a good, you know, series of preventative maintenance, tasks of what it’s actually going to take to keep the machine up and running. And then ideally you can go beyond and get a better understanding about how the machines run in your environment, how you’re specifically running them. Hopefully we can again go back to those sensors, bearings. And again, it’s back to kind of the basics of what we’ve talked about in my thought of, we need to measure and understand how the machine is running. And if something gets very hot, we can say, okay, this got hot.
— Dave Griffith: Maybe we need to check the grease on it. Maybe we need to change barons. Maybe we need to understand what’s going on, especially for those high dollar value items. And then it becomes a, okay, not only are we doing, you know, our preventative maintenance, but you know, we’ve got notifications that we may have problems with X, Y, and Z things. And so, as opposed to going through the process of waiting until it breaks or waiting until we hit our normal preventative maintenance, we’re just going to go through and push up some of those preventative maintenance items. And I think a lot of it is going to be going through the process of understanding, you know, what’s going on with the particular machines again, going and pulling sensor data and other data through. And it, in my opinion, it’s a lot of a mindset. It’s a, okay, we’re not going to be the shop that is going to wait until it’s broken to fix it. We’re going to go out and, you know, prevent some issues. And then we’re going to go continuously, check the machines and see what we can do to stop it from breaking.
— Vladimir Romanov: Yeah. You know, I’m like, Hey, guess on my side. cause I gave this a lot of thought to, based on what I’ve seen in the industry. And I’d be really curious to see if there’s going to be an application where you could, you could pull the data based on cycle times. And I know that’s somewhat done through, you know, some of the high-level like O E metrics, but ultimately what would be really cool to see is let’s say you have a case biker, for example, right So it takes the product and puts them in a box and then sends them off on the conveyor. But I think if you could see that cycle timer to be, you know, a variable that you want to keep within a certain range after your VAT or fat, however you call them. and that slowly starts to decline as the machine wears out.
— Vladimir Romanov: You could essentially not just send alerts, but you could potentially like adjust certain parameters of the machine to see what kind of an impact they would have on that cycle time. And ultimately you could also see like how would the grades over time Cause I think there’s, there’s going to be like very fine adjustments where let’s say, if you’re a case, backer is the bottleneck and it needs to run at like 10 cases a minute. And all of a sudden it runs at like 9.98. You’re starting to see, you know, this bottleneck, but it’s not necessarily like that something’s broken is just like, let’s say your belts are slightly where now your pulleys are maybe not as, as lubricated as they may need to be. Like, your servers are getting, you know, blimp of your network. That’s doing some weird things. But I think there’s a lot of like value in that.
— Vladimir Romanov: And I think based again, like from what I’ve seen, there’s not a lot of that going on yet. And I don’t know if it’s a measure of, like you said, I think, how do you say like attitude towards the process, but I think as you have these multiple assets, it will become more and more valuable because you could find like bottlenecks, you could find issues, you could analyze like what’s going on, what the asset at the, at the production level and ultimately like figure out what needs to be fixed then. Like not only when it needs to be fixed, but like here, you’re running out. Like, you know, it’s like diminishing returns curve, like here, you’re running out like 90%, are you running like 80 and like slowly declining. And like when can you do that
— Dave Griffith: The maintenance I think, yeah, no, I like that. And I think that that should be like some fairly simple math, right If you’re pulling in the information, if you can understand cycle times, if you know where you’re like, if you know where that line is and call it 10, 10 cases per minute, then it’s okay if I get below this and like maybe you hit 9.95 cases per minute, and then it becomes the, okay, the next time we shut down, like we’re putting another maintenance shift on and the cost to run at 11, 10 and a half or whatever. The, the, the ideal, the, the peak run speed is like, that makes sense. Money-wise and I think you certainly can do that. kind of in a less industrial, more real life situation. I actually see that on the van, one we’re driving, we do, we do oil changes like every 5,000 miles, let we get a fresh oil change.
— Dave Griffith: And depending upon who is driving, you know, we can average anywhere between 16 and 20 miles to the gallon. mostly depending upon speed, you know, 65 miles an hour, we’re 18 ish miles to the gallon. But as we get closer to that 5,000 point, we’re dropping a, a mile, maybe a mile and a half, because of the oil. And so before we go for long trips, and almost certainly every time, we’re about 5,000 miles, we go and spend the $50 to get the oil change because you can do the math and it very quickly makes sense to get, I mean, beyond just the, you should change the oil to change the oil, but it very quickly becomes the, it we’re actually saving money by changing the oil. And I think that, that, that’s kind of like the next step, like much of the time, when we, again talk about Bowie, it becomes the, okay, like let’s first time we’re going through this.
— Dave Griffith: Let’s look at the lowest hanging fruit. It almost generally isn’t, we need to do prevent predictive or preventative maintenance. You know, it may be a variety, but most of the time there are much larger, you know, pieces of hanging fruits, in there before it gets to the point of, you know, doing maintenance. But once you have that data mindset of like, let’s look at that, let’s see what the problems are. Let’s be able to drill in. And maybe you have a person or a team or a couple of power users who do a lot of that. That is going to be information that is basically at your fingertips. And again, it should be a fairly easy calculation once you understand the variables. And once you understand the variables, you’re going to ask why I haven’t been doing this for the last five years or 50 years.
— Vladimir Romanov: Yeah. I mean, I think it comes unfortunately from the industry, as Frank mentioned, the, custom machine builders are not as interested in, you know, the longevity per se of that machine versus, you know, the design, I guess, of the initial ones. So I think they cut costs and it becomes really difficult for them to, to create that kind of a maintenance plan and add all the hardware that would be needed, not only hardware, but software as well. Right. Cause ultimately these systems, I don’t think they could reside in like an edge gateway. It ha it would have to have some kind of a server that collects the data that ultimately analyzes, you know, thousands, if not millions of different points over time. Because I think, I guess like, that’s like, I think it’s good that you brought up the car industry, right.
— Vladimir Romanov: But that’s what the, the Tesla, the Elon Musk is doing, right Like he’s pulling a lot of data from all of the vehicles so that in the long run, they would be able to create the autonomous driving or self-driving vehicle. But ultimately you need, I think a lot of data points to get that done. And I don’t know if machine builders get the same volume, you know, obviously they don’t as Teslas, but, enough of a volume to be able to make those very concrete recommendations. So I don’t, I don’t know if we’re there yet, but
— Dave Griffith: So I’ve had conversations with various machine builders, some very large, some very small, and most of the time, those conversations get to the point of we’re going after a low bid, the potential value add that I am offering by adding some of these services could look like even if it’s $5,000 or $10,000 could easily be the difference between me winning this machine build and me not wanting this machine. So the machine builders, I find generally aren’t as interested in kind of going down that path of putting them in as, as an add-on or as a value add, because that’s not their business. most of the times that facilities that I’ve worked at and worked with that are interested in that are the, this is what we’re going to do as a facility sometimes is going to be an add on to the project. Or sometimes they’re going to demand X, Y, and Z capabilities for any machine that comes to the door so that it can plug and play into their system.
— Dave Griffith: And I think it’s going to be much more facilities, by industry, by verticals who are going to say, okay, it is such an issue. We can save so much money. We’re going to, you know, go and push to do this and to kind of hammer on a dead horse. I mean, there are lots of facilities that aren’t even running basic OEE calculations, or being able to go through the process to understand like where they are compared to where they were cycle time or, or other things. And if they don’t understand that we’re, we’re so many steps away from wanting to have some of this functionality built onto what’s a new machines or retrofitted machines.
— Vladimir Romanov: They have a question on YouTube about, your background. I don’t know if you want to give us like a one minute spiel. People are curious about what it is that you actually do.
— Dave Griffith: That’s fine. so Hey, YouTube. my name is Dave Griffith. I am also kind of curious as to, as to what I actually do, sometimes, but, the, the, the long and the short of it is, I’ve personally been in the industry for the last 10 or so years. I grew up in the industry, as well, I’ve done a little bit of everything, kind of from working at, at facilities in warehouses, to working at OEM machine builders, building large facilities, from that point, you know, I worked at a manufacturers, rep distributor did a lot of hardware work. I ran a systems integration company focused mostly on the software, like, MES manufacturing, execution system level. And now I guess the easiest way to describe what I do is I’m, I’m a manufacturing consultant, although I don’t necessarily love calling myself a consultant because you kind of get the eye roll and the, Oh, he’s going to come in and tell us we’re doing everything wrong.
— Dave Griffith: And he’s got to, we got to pay him a lot of money and then he’s going to walk away. I like to help facilities and people, you know, go through those buzzwords, right That industry 4.0, that digital transformation. And generally I’m working on, you know, a couple or a handful of projects, as well as helping, you know, a couple niche service providers, you know, build their, build their offerings into, into the manufacturing industry. I, I, I like kind of understanding where we are and helping people understand here’s where we are. And here’s where we’ll be in 10 years, which is why I’m doing some, some more on these real work goggles, because while they may not be super prevalent today in two to five years, I think we’re going to have a lot more technology on the plant floor. And I, I certainly want to be part of that. and then on YouTube, if you guys want to check me out, I’ve got a personal blog, Dave hyphen griffith.com. I talk a little bit about a lot of stuff. digital transformation industry, 4.0, and, kind of a bunch of my experiences. So you guys can check that out and you’ve got any questions. If there’s any way I can help you, you can, you can feel free to drop me a note online on LinkedIn, anywhere.
— Vladimir Romanov: Sweet makes sense. What, I guess we’re a little bit over time, but I still wanted to, touch base on our fourth topic. Since we’ve committed to this, we might as well go to full hour, but, big data. So big data. Again, I feel like it’s a somewhat of a marketing term, somewhat of a controversial term. what are your thoughts You know, maybe introduce it with what it is first and then give us some opinions
— Dave Griffith: And insights. Yeah, I think big data is basically, and I think big data has been around for at least a decade. And like the concept is we’re collecting data. We have this data and now we’re going to use this data in kind of the broadest sense of terms, in the industrial community, we see a lot of people using your process of story. And so a lot of that is your OSI soft pie. We see some people using other time series databases. we see a lot of influx DB. we see Wonderware has kind of rolled their own or a variety of different, processes of storylines. And so if you’re one facility, you kind of take that and you put that in a place that people can use it. if you’re in enterprise, you’re taking that and ideally pushing all of that up to a data Lake, almost exclusively in the cloud, and then being able to run some information, and run data analytics based off of that, it’s kind of the, well, once we have clean data in a data Lake, we can use, kind of our, our, normal tools, and our business intelligence tools, augmented or no artificial intelligence and machine learning to kind of dig into that further.
— Vladimir Romanov: Speaking of which AI and ML are you seeing some cool applications and seeing some, I guess, projects maybe that you can discuss or disclose with them
— Dave Griffith: I I’m seeing a lot of people talk about it. I’ve been seeing people talk about it for years. I am seeing very well. I am seeing some machine learning that is actually machine learning. I am seeing, I’m not sure I’ve seen in actual artificial intelligence application in the industrial, ecosystem mostly because I wouldn’t, I would say for it to be an AI, it would have to have control. And I haven’t seen anyone give a industrial facility control over to the machine for a variety of reasons, but I’m certainly seeing more people kind of looking into that, talking about big data and, and artificial intelligence and machine learning. I think that some of the biggest issues are the fact that the data, I mean, it’s dirty, right Like we, one, we set up, you know, at the most basic level, these PLC tags, you know, 10, 20 years ago, it was Dave, does it one way black does it a different way. Jim, Tom, Terry, and Tony all do their own kind of sorta within some sort of structure what all slightly different. So it’s very difficult. You have to go through and clean a lot of this in order to put a process in place before we could actually look at, you know, running any sort of machine learning or order artificial intelligence on it.
— Vladimir Romanov: Yeah. And if I, if I may expand on that, I think it’s not only, you know, the differences in programming styles, which of course plays a very big role in getting that data into where it needs to be and ultimately become usable. But also the capabilities I find of many facilities are, as we talked about, just not there yet. Right So I go to many manufacturing floors where a lot of their even PLC systems are not even on the plant network. So it becomes extremely difficult to, you know, to go into those meetings and kind of crush the dreams of, you know, some of the manufacturing leaders when they’re just telling you like, Hey, we’d like to get AI machine learning into place. So though we could manufacture 10 X the number of products that we’re currently putting out. And, you know, they’re just not ready.
— Vladimir Romanov: It’s not even like a matter of getting the data. They’re just not even connected. They don’t have the right instrumentation. And again, this ties back to some of our IOT discussion where they might need to get, you know, certain hardware upgrades, which then come with certain software upgrades, which then come with certain software programming methodologies. And then again, like we can talk on and on, but there’s going to be, you know, protocols that you’re going to be using databases that you’re going to be using. You’ve mentioned a couple. and I know certain facilities, again, like depending on the application would use, let’s say graph database versus like a traditional SQL database. Like there’s a lot of different, I think, like cogs in the machine that, again, may not be applicable and not as generalizable as some of the vendors may make you think.
— Vladimir Romanov: So it’s, it’s extremely important. And I think it’s, certainly headed in the right way. But as you said, like, I haven’t seen any like very specific applications yet. And again, I think a lot of people mislabel what AI is. And I, I liked the fact that you kind of pointed that out because ultimately to a, again, a very non-technical person, something as simple as controlling the level of a tank based on a sensor, it could mean like, Oh, that’s that’s yeah. You know, we’re controlling the level automatically. Well, like that’s not really the definition. So it, it’s, it’s a very fine line. I find between like what marketing says and what is actually delivered on the plant floor and based on the traditional, I haven’t seen something that’s, I guess fulfilling the criteria is for me at least.
— Dave Griffith: Yeah, no, I think those are good points. I have, I have a couple other thoughts on that. so the closest thing that I’ve come to people actually using that to run in line is it’s kind of back to that facility that I’ve been to that, makes aluminum cans. and one of the guys in the middle was telling me they were going through a process to kind of pull in all of their process variables. And so as the item is going through the mill, basically, they’re going to be able to tell the item at any point in time on the line, based upon the process variables, if they continue down that path through like normal running operation, if it will turn out good or bad with the theory that if it’s going to turn out outside of process variables, they can change something as they’re going through that process.
— Dave Griffith: So, but it’s basically like a really difficult we’re linear regression, right And, and like, this is, this might be getting too technical, for this particular stream, but like that’s not machine learning that that’s a, that’s a linear regression that if when theory we could run on Excel five years ago. And so, but, but seeing things like that is, is very exciting. And I think that those guys in some instances are certainly like on the bleeding edge of like let’s take data and make sure that our output is good output. I’ve seen some people kind of do some of that with multi-variant Annette. Well, I’ve heard of some people do some of that with multi-variant analysis in like the pharmaceutical industry. Whereas if they’re running down the tube of known good, they know that the end product is going to be known good.
— Dave Griffith: I think that’s because pharma is kind of towards the bleeding edge and there’s a lot of money in it. And I would also imagine, especially in some of the mixes, like if you screw it up, you can’t melt it down and try again, it’s you you’ve screwed it up and it’s scrap. And then there are probably different, FDA and other food and drug industry regulations that are issues. And the, the other comment, and I feel like I cannot talk about big data or arm without, talking about Jim Gavin. Again, he and I have done a bunch of videos specifically, on this over the last year, and his best thought about this is actually on YouTube, right So YouTube has probably the most data and the most viewing data out of literally everything that we want. And sometimes you say that you want to type fishing into the, you know, I want to see a fishing video on YouTube and it gives you a golf video or a cat video in the top 10. And those guys are literally spending all day every day, you know, running through the process of trying to create a better algorithm to get you to watch more video.
— Vladimir Romanov: Yeah. Yeah. I think, Jim is doing a lot of interesting work. I know that he doesn’t share all of it publicly, but I’ve watched the couple of videos that you’ve mentioned, and I think there’s a lot of insights. And I think that’s part of the reason, I guess, why the industry is moving so slowly is that there’s not enough people to educate about, you know, best practices. And I think ultimately, you know, different sectors for one reason or another, I guess confidentiality are unwilling to share, you know, some of their findings. So even if some manufacturing plant is leveraging AI or machine learning for, you know, I don’t know their recipe, utilization, for example, they’re not necessarily going to want to share that with the, with the rest of the world. Right. So I think that’s part of the reason why also, since it does give you an advantage, it’s, it’s really hard to kind of find these things, but as you had mentioned, I guess, some of those applications, one came to mind.
— Vladimir Romanov: I remember hearing about a chemical facility where they had used a digital twin, which is, I guess, like another topic of industry 4.0 that we can discuss down the road, but they’ve used that in conjunction with, you know, the batches that they’ve collected over over the years. And then they were able to reliably kind of predict if they were going to change the mix, how that batch would go through the process. Right. So again, like, it’s, it’s really a multivariate analysis, but I think they’re still pulling historical data and kind of like analyzing that. So I think it’s at the basic level of, like machine learning, but it’s still kind of is there, but that’s one of the use cases that I’ve kind of heard of never seen in person, but you know, something to kind of watch out for.
— Dave Griffith: No, I think that’s very interesting and I’ll tease it cause I know it’s coming out soon that there’s some interesting, multi-variant analysis videos that the Jim and I shot when we were together coming out, on the industrial insight channel, I think end of February. and when it goes live, I will, I will mention it on the podcast. and I think it’s, it’s very interesting. We got pretty deep into a couple of different, applications that I think people are going to want to watch. I know it was mind blowing for me. I had heard about it twice before we, before we sat down and walked through it in person. And I think I just sit there with like a stunned look of, wow, we can actually do that. That is so amazing. right now,
— Vladimir Romanov: I mean, I think it would be great to have Jim maybe on the podcast, down the road, but, you know, if we want to, I think let’s close off the, the current episode, with, you know, some final couple of words and I guess I’ll, I’ll start off by saying, you know, thank you everyone. Who’s been watching the show, we’re going to be posting this on all the podcasts kind of mediums, but you’ll also be able to see it again on YouTube, Facebook, LinkedIn, and, and Twitch, I believe saves the recording as well. There’s going to be links in some of the things that, I guess how to connect with Dave, how to connect with me, we’ll add the links that he mentioned about the videos that Jim has created with him together. So that’s going to be all like into the descriptions and the comments. I don’t know exactly here, there, maybe somewhere, but there’ll be posted for sure. appreciate all the participation. Appreciate the questions. what are your, what are your thoughts, Dave Are we ready to commit to the next episode Is that, is that maybe what I’m sensing or not yet Yeah.
— Dave Griffith: Yeah. I mean, so again, I’d like to kind of mirror of lads. Thoughts is thank you guys. All, anyone who’s managed to, I to stay here through the entire thing or even part of it, it, it’s always great to get the live comments. if you guys are listening to us in a non, if you, yeah, if you guys are listening to us in the podcast, please come watch the stream. And almost anywhere that you can watch a stream at this point, w which is pretty awesome. I know Vlad and I both feed off of that and we love answering your questions. I feel, I feel good to, to commit to the next stream. I think we were talking about doing it in two weeks, Wednesday, the 17th. Is that a, is that what you’re thinking
— Vladimir Romanov: I think we’ve, we’ve wanted to commit, I guess, to one day of the week. And I think Wednesday really works well for us as well as many other viewers that we’ve kind of asked that question. And so this would put us on to the 17th, right February the 17th would be the next one.
— Dave Griffith: Yep. Wednesday the 17th, six o’clock East coast time, which is noon somewhere. And at some point I’ll figure out what time zone is Noonan and, be able to say noon in that time zone and really confuse everyone.
— Vladimir Romanov: Perfect. And I think, I don’t want to give, give out too many details or like confirm our guests before we spoke to him, but we’ve got some really exciting people lined up. And, our goal, as we had mentioned in the last couple of podcasts, was to bring some subject matter experts that would be able to talk to some of these topics, even more in depth than, me and Dave would. And so that would make the stream even more interesting. So feel free to come up with questions, make sure that, you tune in the 17th of February and we’ll see you guys at that time. Thank you everybody again and see you in two weeks.
— Dave Griffith: Perfect. Thank you everyone.