Transcript of MES Systems: The Secret To Manufacturing Success

why it is that companies ultimately choose to build their manufacturing execution system take [Music] zero all right so in this video in this week's Whiteboard video um which is a follow-up to the short we did on Monday which is a followup to the many videos we've done on Ma manufacturing execution system manufacturing execution systems today what I'm going to go over is what are the technical reasons why manufacturers end up choosing to build their manufacturing execution system rather than buying it off the shelf what are the technical reasons okay so a little bit of background so before we get started a little bit of background if you want to know what is manufacturing execution system we shot a video about that about a year and a half ago you'll be able to get it somewhere up here it'll be linked we shot a VI videos on what is manufacturing execution how to calculate OE and why you need to uh this is what you must know before you build your manufacturing execution system and how to implement OE in your manufacturing plant that's just four of the better videos that we've shot on Mees I highly recommend you watch at a minimum the what is Mees video okay but a little bit of background okay manufacturing execution systems manufacturing execution system is essentially the place where we convert a um something we've sold in our Erp system into manufacturing so if you look at our manufacturing work flow okay if you look at the manufacturing workflow we sell stuff in our CRM we plan to manufacture it in the Erp the Erp is sort of the master data model for the business it's the place where we store all the products that we sell the definitions of those products the manufacturing steps the inventory the raw material all that kind of stuff okay but Erp is really planning software it is an execution software the manufacturing execution layer the mees layer is the place where manufacturing actually takes place okay it's the software you generally see on the plant floor or in the supervisor's office actually on and okay if you if you were to look at CRM Erp Mees scada PLC HMI warehouse management system shipping an ARP and you were to if we were to put this in terms of Representative sizes so that is which of these pieces of software are the biggest okay it's actually going to be manufacturing execution in most organizations and and even if it's not um software let's say it's just paper execution okay the the the traveler the work order the barcode that we scan the raw material the the sheet off the raw material from inventory etc etc etc manufacturing execution is going to be the biggest of all of them okay Erp is going to be huge but it's sort of going to be the tip of a pyramid and underneath that pyramid is going to be many instances of manufacturing execution systems in all of our sites okay so manufacturing execution is absolutely the most critical piece of software you have in your business there's no question about it it's the most critical piece of software okay so what we're going to be talking about today is why is it clients end up building their manufacturing execution system rather than buying it off the shelf the way that most people start is they they call Rockwell and they say we want Rockwell Mees or we want wonderwear Mees they go to wva and they say they want wonderwear they they go to seens and we say Seamans we want your Mees software they go to Traxxas and they say hey traxxis we want your Mees software we go to PTC and we say hey PTC we want to use your manufacturing apps or your control apps for our manufacturing execution system okay they they generally start with off the-shelf and what they discover pretty quickly is that the cookie cutter uh software that they bought off the shelf doesn't fit their business okay and today I'm going to explain why the technical reasons why you end up building all right all right so if what we want let's first talk about what Mees is and if you look at the what is Mees video one of the things I talk about is the core four okay so that is work order scheduling overall equipment Effectiveness and downtime tracking those are basically the four capabilities that are in essentially every manufacturing execution system okay and then there's a whole Suite of other capabilities ities you're going to pick based on your business needs for example a really good one would be like recipe and formulation let's say I'm in a batch process okay and I'm mixing some type of batch in a tank and then I'm going to a linear discret process that's chopping that that bulk thing up into smaller discret components the recipe that went into making that batch and the formulation all the the lot numbers and the the individual items that went into making that batch are critical to the performance of the manufacturing later on Downstream but that doesn't always apply that doesn't apply in every manufacturing process so if I were to buy my Mees system off the shelf and it wasn't specifically designed for batch processes I would be out of luck if I wanted to be able to tie the uh the recipe the formulation and the batch number to my Downstream manufacturing operations okay all right but in general customers want something like this at the line level for the for the the op the general operator they want to be able to look at certain things okay so the general operator wants to have insight into what we're running our work orders okay that's this upper leftand corner here in general they want to have insight into how well we're running okay and this is your oee this is OE by run down to availability performance and quality and this is oee by shift then they want to have maybe maybe they want to have some insight into how the individual cells within this production line are running so the same view here except now we're looking at the sub the subunit so if this production line is made up of say four cells what I want to be able to do is select between my four cells to see what these these efficiency numbers are by individual cell okay and then our downtime tracking okay in general in general want to build something like this for our operators okay and this is your generally your line level dashboard so this is this is visualization number one that we want to see visualization number two is we want insight for the supervisor so a supervisor is going to be responsible for more than one production line generally supervisors are in charge of an area in a manufacturing facility or many areas in this case we're going to say that this supervisor is only responsible for um these four production lines okay they want to look at the schedule and they want to look at the individual production lines then what we want is we want to be able to do analysis okay so what we want to do is be able to predict a select a production line and what we want to do is we want to look at what are the top 10 causes of downtime by our production line by production line and then by our individual CS we want to be able to do analytics okay and then we want to from a supervisor level we want to be able to look and see how did we run over the course of a shift so all of our individual cells our production lines when were we running when were we not running what was our production relative to what we should have been producing the dark blue bar is what we produce the light blue bar is what we should have produced actual versus Target production defects by individual cells this is top five uh line downtime reasons Paro chart and these are all the individual lists of downtime events okay we want to build all those things that that's our core th that what I just showed you there is your core for Mees okay so that's the core for that's the thing that everyone wants in their manufacturing execution system there's no additional features there there's no recipe and formulation there's no quality management there's no batch tracking um there's no statistical process control there's no statistical process analysis there's none of the there's no bomb kitting there's nothing special about what I just showed you okay and yet you could not have built what I just showed you right there with an off the shelf the solution for that specific customer and that specific customer that I just showed you is a incredibly standardized manufacturing operation and yet you still could not have built that manufacturing execution system with something off the shelf what you needed to do was buy an iot platform a software platform and that's designed for the industrial Internet of Things what is a platform that's designed for the industrial internet of things it's a piece of software that can connect to Industrial Hardware and software it has the connectors to connect to all those things and then it's an environment for doing data operations converting the individual events from all the things it's connect to into information information modeling and then dashboarding and Analysis that's what an iot platform is it gener has a SC scripting language it generally has an engineering environment where you can build visualizations so you want to take an iot platform and you want to build your manufacturing execution system but let's pause for just a second let's say if I was going to build an off-the-shelf Mees system how would I do it okay so I would start I would start by trying to build like every other software is is built okay back endend AP I and UI this is the part that the human being interacts with okay this this API might have we may have uh connectors here that connect out to our machines okay and the API inst stores events into the back end how do I how would I build a manufacturing execution system well what I would do is I would basically create an in an Ingress Point okay which uh let's say I'm going to create an object called a cell and it's going to be in the shape of a triangle okay so I'm going to model what a cell is and I'm going to put it in the shape of a triangle I'm also going to model what a line is and I'm going to put it in the shape of a square and I'm going to model what an area is and I'm going to put it in the shape of a circle okay so I'm going to have an information model in my back end one for a cell one for a line and one for an area I'm also going to write methods in my API methods in my API that can call the triangle call the square or call the circle and then I'm going to build visualizations that will display the cell display the line and display the circle now that's all fine in Dandy when you're building building software when you're building software this is what you want to do okay this is how software Works here's the problem here's the problem the problem is not in defining what a cell is and the problem is not defining what a line is and the problem is not defining what uh an area is the problem is in defining the functions underneath the cell functions like give me all downtime by minutes give me all uh give me oee by cell show me the availability over the last 30 minutes in order for me to create these functions what I need is standard data ingestion I have to be able to model the data the individual events so that they can be ingested the way that that triangle is designed the problem is is that if I look at a manufacturing facility um a cell is not a cell is not a cell the May a cell is structured here this cell may have state and it may have uh it may have a line state Register this one may have just a list of alarms this one may have just it's stopped or it's running that's it okay so what am I going to do how do I take how do I normalize a cell that has a state Register that is a number if the number is 31 it means it's stopped for this reason if the number is 44 it means it's stopped for that reason if the number is zero it means it's stopped if it's one it's running that cell has that model this cell just has a list of alarms so I could have 500 alarms 31 of them are either active 31 are active and the rest are not and this one just has a one or a zero one means it's running zero means it's stopped okay most of this software in in fact nay nearly all the software that comes off the shelf isn't going to work for all three of those machines machines what you're going to have to do is make those machines fit the software the way the software was designed the mees system off the shelf was designed to consume the data so um a a really good example of how this works in practice let's say that you call up um I'm trying to think of the companies that have oh there's a there's a Mees system called plant forance that is um I I've been seeing around it's a French thing uh that plant forance um Mees system basically has an edge device okay and that edge device is designed to essentially normalize that data so basically if I I'll put my Edge device out on the plant floor okay and what that edge device is designed to do it's designed to capture the stuff from the machine and format it the way the triangle the cell and the area needs to consume it and the way the functions go get the downtime for the cell go get the o for the cell is designed to process it uh I think Corso systems does a similar thing they have like a a little box right and they put that box out there that box is designed to do this thing in the red take the stuff from the machine and format it for the way their system is designed to consume it okay that's problem one problem one is is what happens once you install this Appliance or once you once you integrate the software with these m manual connections to our functions well what happens if I want to make a change I want to move a piece of machinery for example or I install a new piece of Machinery or I update an existing piece of Machinery that breaks this what do I do there okay that's problem number one but the bigger problem is is since manufacturing execution systems aren't it's not a solution what it is it's a list of capabilities that's basically starts with the core four go ahead and watch the what is Mees video Work ORD scheduling oee and downtime tracking and then we just add a bunch of other what if the off-the-shelf software doesn't have the other or the four things that we want what do we do the answer is you're out of luck that's why people end up building people end up building because of this and because the limitations and the list of capabilities so what does that look like okay so what does that look like in the end what you end up doing is you get a platform right so I'll end up with my platform in the center in this case let's say it's ignition ignition is an iot platform that on on one side it can talk to all the stuff on the plant floor okay talk to all of our machines okay on the inside it can do the three things we need it to do so it can do data Ops now in this case we wouldn't generally do the data Ops we would put we would use highb for this layer okay but it can do the data Ops all right it can do the info modeling and it can do the U I and Analysis and then on the other side what we have is our standardized visualization of our triangles our squares and our circles and our X's when we decide to add an X and our um y's when we decide to add our y's that's the technical reason why people build manufacturing execution systems in order for me to get this in order for me to get this for my operator that for my supervisor that for whoever wants to analyze the reasons why we're not performing our availability number isn't as high as we wanted to be or our quality number isn't as high as we want to be of our performance number isn't as high as we want it be whoever needs to consume the data analyze the events in order to build that or in order to build this where were we to plan in real time what was our production this hour relative to what it should have been where are where are we trending over the course of our shift more over this is one of the reasons I love flow software which is an iot platform one of the reasons I love slow flow software and the the unified analytics framework which is coming out from the team atlow is that we can take this data here and we can use linear regression to to predict the future Midway through a shift or Midway through a month or Midway through a week or Midway through a quarter okay in order to do all these things okay either you have to buy a piece of software that has the ability to ingest what it is you have on the plant floor okay and format it for the objects that they've already pre-created or and and provide you the features that the capabilities that you're looking for or you do it yourself one of the advantages of using an iot platform especially one that has an SDK or has thirdparty partners that develop modules is that you can buy the iot platform you can install a module that does say this piece so this would be say sepasoft Mees modules which we don't use okay but what sepasoft module does in this o iot platform is it can consume the data from the machines in the structure that the sepasoft analytics engine needs to do your calculations and then you still have all the flexibility of the iot platform to extend the capabilities that sepasoft modules can provide for you okay which is what I just showed you there is actually an example of this architecture but in order to do this and what we've been teaching in the core manufacturer the core Mees boot camp which started last fall and actually ends this Saturday and what we'll be doing in the advanced Mees boot camp is we're teaching you how to actually do that so when I was I tweeted this out yesterday this is the process of consuming raw asset data whether it's via OPC flat file mqtt or it's manual events from operators okay you have line and cell data abstraction okay so what we have to do is we have to take the events from the assets we have to for for format them into specific events with definitions the most the four most important ones for the core four are in feed out feet waste and status and then we can also add in custom parameters and attributes from that space this is actually a name space then we then map into our manufacturing execution data and information model so that's where we consume um events to do specific functions like calculate oee like calculate availability like calculate performance and quality okay and then what we do is we consume that information from those models and we analyze it in visualizations this is what we've taught in the core Mees boot camp we taught from a theoretical level and then we built it and then this Saturday what we're doing is we're revealing all the stuff that we built what we're going to be doing in the advanced boot camp the mes boot camp is showing how to build but we're going to be building extended features specifically the big one will be Erp integration if you watch the what Mees video what is Mees video one of the things I talked about is if you look at the core 4 work order scheduling uh OE and downtime tracking the very first one of those functions that gets dropped is scheduling it's literally the first one that'll get dropped the second one that gets dropped is work order management okay what we're going to be doing in the advanced Mees boot camp is showing you how to consume work orders from Erp how to consume a schedule from an external third-party application without managing it inside of your Mees application but I can tell you this I've been doing manufacturing execution systems for a very long time okay two major observations that I've made one about the systems integrators that do it and one about the end users okay the systems integrators who are off-the-shelf people they generally only work with one piece of off-the-shelf software right they they try to Master One Piece of off-the-shelf software and and they try to find the clients that work best for the capabilities that off-the-shelf piece of software can provide and they go in knowing the limitations of that software and what it can't do and what they try to do is just steer the customer towards what it can do then you have the integrators who build and the integrators who build always try to do the same thing the vast majority okay what they try to do is they try to do it the way the off-the-shelf people do it they start with a specific list of capabilities and they'll build some type of edge device that they want to sell to the customer for for $1,000 a machine or whatever it is and all that edge device does is connect to events and turn those events into a shape and then connect that shape to their shape that they already have in their software okay the best way the observation about the end user excuse me the end user is never satisfied with either of those outcomes as the end user gets smarter and smarter and smarter during their digital Journey they become less and less satisfied with the limitations of those two approaches which is the reason that manufacturers always end up building always and I would say that if you look at the vast majority of manufacturing execution systems we built we're not going from paper to digital based on an iot platform unified namespace digital strategy architecture based on EDG driven report by exception of lightweight technology we're not going from paper to built Mees we're going from off-the-shelf Mees to built Mees um and that is those are the technical reasons what which I've just laid out as to why it is people end up building all right that's my story I'm going to stick to it like subscribe uh click the notification notification Bell share the video uh thanks for watching and I will see you in the next one