<
and with me today, i have three panelists. all of us are part of a coalition known as the 'smart manufacturing leadership coalition'. and so, we're really here to represent the work that's ongoing with that organization. and so, just briefly, i'll introduce our panelists and then i'll introduce myself, give a little foundation of how the smart manufacturing leadership coalition' defines smart manufacturing. and then we'll jump into conversations here with our panelists. so, first of all, here on my right is steve prusha. he is the manager of strategic systems office at the jet propulsion laboratory, the california institute of technology.
steve is responsible for the development and implementation of new advanced engineering environments and design methodologies, analytics for assessing and characterizing new complex systems and technologies, and risk modeling, and management capabilities supported by various sponsors at jpl, nasa, and other government agencies. he has over 30 years of experience in technology development, spacecraft design, flight system development and operation, model-based systems engineering, and project and program management at jpl, nasa headquarters and industry. so, this is steve. beside steve, we have lance fountaine. lance is the industry principle for metals and mining industry at osisoft,
following a 20-year career in the aluminum business with alcoa global primary metals. in his last assignment before leaving alcoa, lance was accountable for the global development and deployment of common best practices manufacturing applications, as well as a supporting computing infrastructure. the renewed focus and resulting strategy led to the adoption of a smart manufacturing program across the global enterprise for alcoa. this program was based on the adoption of the pi software system from osi, as an information infrastructure to support efforts for continuous improvement, operational excellence and ongoing business sustainability.
and down on the end, far in the other end, is mark goodstein. mark is an entrepreneur-in-residence at idealab, a business incubator in pasadena, and a co-founder of systemex, a start-up that sells systems engineering tools to manufacturing companies. mark is an accomplished start-up veteran, a start-up adviser, and public speaker. he started several successful companies and a globally recognized competition to spur the market for high-mileage vehicles. mark serves on the la n sync bell commission, a group of civic leaders for the los angeles basin,
supported by the ennebelth and burke foundation, who want the region to begin claiming its equitable share of federal state and foundation dollars. he also serves on the board of innovate pasadena, an organization whose mission is to catalyze the innovation community in pasadena, california. that's right. three panelists, i myself, as i've mentioned my name is mike yost. i am the president of mesa international, i'll give you just a very quick introduction, who we are, and then and then give that definition of how the smlc, define smart manufacturing. and then, we will jump into it, talk about it as a group, and interact with you as you see appropriate. as mentioned before, i represent a non-profit industry association.
and our role is to help educate the marketplace on how and why to use modern information technologies in manufacturing. a lot of the previous conversation from dr. petrick and a lot of the things that we see in our daily lives show that the speed with which technology is coming at us. i can tell you when mesa was founded in 1992, technology was more likely to hold us back as manufacturers, as far as getting into the admission critical applications. and today, we're obviously on the opposite side of that, where technologies are exploding and moving at a pace that is very difficult to keep up with. so our role within the industry, is to work as a global community with manufacturers, solution providers, consultants. get the community together and focus on the business value, as well as the tactical ways and best practices for implementing technologies. so with regard to the smlc itself and smart manufacturing,
the smlc is also a not-for-profit organization, which has its origin with the national science foundation initiative that has been working for several years to establish a road map for using advanced process manufacturing technology in manufacturing or what is known as smart manufacturing. and if you can see from the slide here on the one side, there's a list of the types of partners that are there and you'll see academia, industry associations, solution providers, manufacturers, everybody coming together with the goal of driving an industry, driven open architecture shared infrastructure. so, what we call a smart manufacturing platform. and we will talk about that in a minute and throughout the course of our conversation.
but the real goal behind how the smlc defines smart manufacturing is to make real-time information available when it's needed, where it's needed throughout the entire manufacturing ecosystem. so, how do we get access to the information in the context we needed. so, when defining smart manufacturing holistically, as you can see here from the slide, it's a sophisticated practice of using data-driven manufacturing intelligence. how many folks have heard of the term ""manufacturing intelligence""? anybody have a strategy or systems in place to use manufacturing intelligence in your business? alright, not uncommon. the goal as we define smart is to use... it's easy to say we should work smarter. the question is how do we actually go about doing this?
so, if we can capture the data, we capture the information that's coming out of our facilities, it's in our supply chains. how can we use that to drive our... a couple of things that are listed here? deeper behavioral understanding of the processes through modelling and analysis, as we talked about previously. the ability to take action, as we see things happening, as we have the infrastructure in place to understand where problems are. anticipate problems, attack those problems. new insights for innovation from a broader base of innovators, which is something i believe personally is a very big upside for what the smlc is proposing. and then, we'll talk about the shared infrastructure, the ability to reuse that. it have capability from all these different innovators, in a way that dr. petrick talked about the it capabilities that need to be... that will be a must in the future.
the focus with the platform is to help enable that without putting all of that responsibility back on each individual manufacturer. so just... just sort of plug through this with the factory at the center with access to that information as we've talked about with that infrastructure in place, it really enables the ability to do a number of things. plantwide optimization, and agile demand-driven supply chain, sustainable production when you have access to the information, when you have that information available to you. that gives us the ability to really solve any of the problems that we need to solve. so, this is the vision that the smlc is looking to drive and the way that they're looking to drive it is again through the creation of this open platform. and we'll see here also, the talk about manufacturing application store, an app store and you see in the bottom picture, the ipad. very familiar environment.
people used to buying apps, dropping them in, and using them. same sort of concept applies there, that we can have more applications working to solve more problems but in a more integrated way. when you buy an app from apple, you don't worry that it's not going to work, you know it's going to work, you drop it in and you go. that's not the reality in manufacturing when it comes to applications from that perspective. so, this is a depiction of some of the work that is going on from a technology perspective. and we're not going to get into this. anybody wants to talk about this, we can talk about it offline afterwards. come and see the folks over at the table in the other room, we can talk a little bit deeper. but this is just some of the work that the team is diving into, to provide that core infrastructure. and we'll talk about what the value of that is for you. but to go back to some of the earlier points about the need for it-savvy capabilities moving forward.
by offering a platform like this, by offering it as an industry open platform, the goal is that you have less of the requirements to have all the people in your organization that are capable of knitting together things like these for your organization. and more so, focus on what the business value is, focus on the processes and then find the right way to enable the business processes through the technology that you need. so, we talked about everything from the data collection layer to contextualizing the data, having internal tools and the app store and basically providing that internal infrastructure. and i'm going to end with this... on this slide. this is from general mills.
now, this is a slide that they provide, that basically maps out what they call their eco system of ""stuff"". and all of these applications that are listed here are core systems, and core functions, and core applications that they have within their business. and this, in my experience is about as clean a look as you're ever going to get at somebody's hierarchy, their ecosystem, and their understanding of what they have and what they need. most people have no idea how to get it to this point and they're one of the primary drivers behind the smlc. they're very active in this. they see the need to have this infrastructure platform because most of those applications are either from various different vendors or homegrown themselves. and so, just to be able to support the core infrastructure to support all of these things, is significant and it's worthwhile for them if they can get out of the software business, basically. they can focus on
food and their business problems. so with that, that's a high level definition of what the smlc is about. what's smart manufacturing is defined as from our perspective. so, i'm just going to pause for a second and ask my panelists if they have anything that they would care to add to my definition, anything that i glossed over too quickly. <
but i think that was reflected very well with really what you said. <
that they made, that they could not possibly roll that out to their supply chain, that made that want to be involved in the smlc. in order to build an open computational platform for others to participate in and be a part of their sort of global presence of food production. i think that's a key point in driving what smlc is trying to do in building this platform. <
okay, so, i think information has a place in how you operate. and if you think about how information's consumed, i'd basically break it into 3 categories. we use it to drive action here and now, we use it to do analysis and why do we do analysis? we do analysis to learn what we don't know today. and then finally, and it's important to a lot of especially larger manufacturers, we drive reporting requirements, whether they be epa reporting requirements or whether they be manufacturing performance reporting. we drive some form of reporting added data. so, again, data and information is very critical to all three aspects. and all three aspects are really important parts of your operating system. <
in particular, in terms of increasing the efficiency of our operations, i'd say, really, we reduce uncertainty, so as the design now is in essence understood and flushed out deeper and broader earlier. okay? we reduce the uncertainty, or at least, better, really, more accurately can characterize the uncertainty much, much earlier. and so, we reduce the number of surprises, that's really what comes downstream. and at the same time when we're doing that, we increase the agility. so in our world, our trade spaces tend to be exceptionally large. we deploy in highly uncertain environments and so there's a premium place down the resiliency of our designs. so our systems tend to be very broad, in consideration. and this is where it's really helped us. it's simply a flush of stuff out earlier, to move from point designs to trade, to a large number of point designs to trade spaces and trade surfaces. be able to exercise those long before we have to cut metal and actually go into production.
we have long life cycles. our space craft may take six, seven, eight years to design. three or four more years to build and sometimes they can take 10 or 12 years to actually reach the destination. so, we're dealing with very long life cycles and this makes a huge impact and we're running out of time really, for our kind of designs. <
is in essence delivering their designs in the same way that we construct them, that is model-based form with the kind of characteristics that we need, with the ip embedded in those, up to the supply chain, and the delivering parts as a result of that. we don't have that yet but we're working towards that. the acquisition system as much as anything else, has got to change for that to happen. but we see those changes occurring. and so, the idea or the time when our systems are flushed out in virtual space, if you will, and then the final physical article delivered as a validation of that model, it's common, but we're not there yet. but that is the hope. <> my perspective is different than big company or jpl but i tend to see things in terms of the... and we hear about this a lot, democratization, the maker community, the 3d printers we've heard a lot about today already.
there are... the implications of this platform are to free individuals or companies with expertise to be able to build these models that can then be adopted by other companies to do things. so, engineers working at chemical plants for 20 or 30 years, can now go off on their own with their own knowledge of computational fluid dynamics, build their own algorithms, their own models, sell them in the app store, that was mentioned in the previous slides, and be in business for themselves, right. and there are a lot of problems with that. no company is going to take a model off the app store unless it's been certified, in which case no one will lose their job for having adopted that model. so there has to be some way of certifying that the model does what it says it will do, and so forth. but that's sort of the beginnings of a very exciting market place for small companies, small individuals, people believe, etc.
so, this is really sort of a bottoms-up set of disruptions that we're seeing happening already. and that will continue to roll, as we move forward. <
they may have their own databases, their own data collections, their own communications needs. and then people try to stitch those all together. and i think from an innovation perspective, what i see as being most compelling and exciting about this is, exactly what mark was talking about. in that, when the infrastructure pieces taken care of, then you can focus on the business problems, the domain expertise. we can put people in the business, not that they necessarily will all go into business and start selling software in this space, but that's how pretty much every company that i've worked with has gotten started. somebody with domain expertise and *inaudible* and started to write their own software, do their own thing, and take it out into the market place. the problem is, it's not scalable, it's not repeatable. and unfortunately, that's where a lot of manufacturers are stuck. you got one person or a group of people that are cobbling things together in excel, or patchworking something together.
so, when the infrastructure is there already, and they can focus on using the tools that are already there. they can find a model that works for them and they can find an app and see if it works and test it out. i think we'll see a lot of innovation come from the market place that this kind of platform will enable. <
so again, i think, we have to change our whole mindset. think about back to that operating system. if we make information consumable to everybody, if people that do the job everyday can look at that data and leverage that data to improve the operation, whether they're an engineer, whether they're an operator, then they become innovators. and i think one of the biggest challenges we have, is how do we broaden that base of innovators, stop calling an innovator somebody that sits in a lab somewhere and does product development. innovation means a lot more than that. i think we have to broaden our horizon and our definition. <
this gives us an opportunity now to really broaden and exercise the ideas much more quickly. so to us, it's all about agility. it's about being able to take an idea and making them... different ideas and making them consistent. to broader community for ideas outside of our own lab or even out of the government. and to be able test those objectively and efficiently in a larger system context. so the idea of whether a model is representing an app, or an idea and an app, or something like that. we now can string them together. we can exercise them in a trait space way faster. much better, much deeper, much quicker. this is the real pay-off to us. it's better design knowledge earlier, better agility. and i think this is really where it's going to play out. i mean, it gets back to irene's discussion earlier about... in fact, combining on one side, it was the increased ability to do simulation.
we are clearly in that game. but combining it with this single person... the capability of one person in a garage and how to develop these ideas and inject them somewhere other than their garage, and this is the exciting part that's really, i think, is going to pay off. <> person in the garage who's designing something with inventor, or something like that, ought to be able to do the mechanical engineering, the design, etc, the surface, the geometry, and should be able to find out very easily. and this is being done, to a certain extent, in some places, where they can make that thing. so, do they want one, do they want 10, a thousand? there should be options and direct contact to manufacturers in the area, that can be their fab, at whatever stage they are. but they should also have the ability to get information that is not part of their expertise. so, electrical and thermal and marketing, and so forth.
and again, some of that is being done to a certain extent, but not in any kind of integrated way. and again, since we're talking about this open platform, the idea of this platform is to allow for that exactly. <
you know, that's not really an effective process. so, in my mind, it's about providing a foundation. and that foundation is information. building a model, a production model, a business model, a process model, that really outlines the critical data. and allows everybody to access it and to consume it. today, when we want that data, it's an exercise to go and get it. it might be in a process control system, it might be on the web. and here we are, we're going searching for it. and it becomes an exercise in and of itself to go get the data. we need that data readily available. <
we can buy equipment, or we can buy machinery that actually takes away some of that infrastructure requirement. and now, starts producing data that we can capture. <
and those sensors can be used to determine whether the stuff they're making is within constraints of quality. the last thing they want to do is ship product into the supply chain that is defective in any way, for a big company like a gm that makes supply chain decisions on a daily... sub-daily basis at every plant, if they were to know that widgets coming off the line, two steps down the supply chain, were defective, and be able to stop those boxes from even entering the supply chain, that would be a big deal for all parties involved. and the platform ought to be able to make it available. and moreover, the algorithms used to determine whether these parts are defective or not should be able to be swapped in and out by new people writing better algorithms to make use of that data. that data, in order to allow for those people to develop those algorithms, needs to be available so that people on the bench can be developing and testing real world data. so, the data exists. it's much more of a collaborative issue that needs to be resolved.
how the data is stored, where it is, etc. but again, it's happening. <
we have an antiquated acquisition system, particularly in government that has got to change or go away. we need to be able to request these things be made for us in this particular way, rather than specifying in a four-inch document that then is returned with an 8-inch document and reviewed by a hundred people. like, i want to send a simulation environment out. i want a model back. i want to exercise it with a bunch of other models. and we want to make an award, we want to buy something. and it's common, and it's there in some aspects but that has to change, i think, as much as anything else. data is a problem. as much as the availability of it is, in some cases, were inundated with it in trying to sort it out, particularly... the particular problem that we have is the integration of disparate data sets or disparate models. it's one thing to put together two cad models and a thermal model, it's another thing to add cost modelling and programmatic and implementation and environmental and everything else.
but you got to do it, that is the definition of system and we haven't perfected that yet. but again, that's really the process. <
who has moved on and smart's not good enough now. they're talking about brilliant manufacturing. so, if you guys aren't prepared for smart, now you got to be prepared for brilliant, as well. and that just unnerves me to no end because we're never going to solve a problem, we're never going to stay stationary enough to fix anything and do anything. maybe i'm just too old and i'm not innovative enough to do that but we've got to stop that kind of nonsense and actually focus on the problem. and i think the thing... the other thing that i rally the most around, when i'm out talking to folks, is that we need the manufacturers to be a voice in this. oftentimes, i pull out study after study that says that the manufacturers are taking a 'wait and see' attitude on where these technologies shake out. and chances are they're going to shake out somewhere that you don't like, alright. and so, if you're not driving the solution providers to make the technologies work the way you want them to work,
then you're going to end up stuck with what you get. so, i think we have about five minutes left. and we can open it up to some questions. <
i don't know any company out there that can be everything to everybody. again, with the company i work for now, osisoft, we have this concept of an information infrastructure. but that information infrastructure doesn't mean we're the only tool in the toolbox. that means there're a bunch of tools you can choose from. you need to choose the ones that are best for the specific need you have at that time. and i know your need here, in this specific manufacturing sample, i think it's a lot more related to product design. so i'll let you comment on this. <
the answer is, we don't know yet. and what's filling the market from the bottom-up really, are open-sourced solutions at this point. so, i think it's going to play out that the question won't be who's going to be the software provider, but in fact, who's going to find a better way to integrate very quickly the highly disparate sets of these models and model packages that are coming out. and i think that's where the answers are coming from right now. it's going to be... so in other words, what it will look like from the standpoint of us as a buyer, for example, is i don't care how you get it to me, and i don't care the format, i don't care whatever, i mean it's going to have to be something i can at least characterize. i'll find it when it integrate. and right now i can tie together almost anything. i mean, i can tie excel to almost anything. and if you can do that, you know... it's... but the mathematic... and i mean, the open-sourced stuff is getting so good, and so cheap and so fast.
i think frankly, the plm folks, if they don't get it together in the next few years or so, they're gone. but it'll be there. <
international conference on additive manufacturing & 3d printing,there are a hundred tools, a thousand tools. but it's all about the data, pulling the right data together to solve your problem.
<> we've seen technology being developed at zeroxposur that does specifically what you're saying. autodesk and 3d systems have been doing things in this area. there are a bunch of other companies that are independent that are trying to solve that particular nut. it's active.
