We help manufacturers Digitally Transform. Learn to lead Digital Transformation initiatives, architect, and build Industry 4.0 solutions using IIOT and Unified Namespace (UNS) concepts at iiot.university.com.
Subscribe! 👉bit.ly/SubTo40Solutions
Follow us on LinkedIn! 👉 www.linkedin.com/company/4point0solutions/
Join our community Discord! 👉 bit.ly/industry4discord
Take the FREE IIoT Mini-Course here! 👉 bit.ly/iiotmini-course
Join our Mastermind Program for Digital Transformation leaders! 👉 www.iiot.university/digital-mastermind
I'd suggest that the significant difference between these Basic Programmable Control Systems (BPCS) is in the way in which the PLC and DCS are programmed, and this stretches back to their respective geneses. Using a programming "language" analogy, PLC programming, even with the implementation of IEC61131-3, is effectively machine code. It processes very fast, perfect for factory automation situations and simple PID control, but it takes a very long time to implement the type of complex computational code to the degree required by process automation, since everything literally has to be "hand-crafted" from scratch. DCS on the other hand, is akin to programming with a higher-level language. Analogue values just are analogue values once an analogue input has been instantiated in the code. There's no faffing about with bits and signing, like there STILL is with PLC code, even when working in the closest 61131-3 implementation by way of Function Blocks. With most DCS programming/configuration, it's just jump straight in and start manipulating real numbers in code as one wishes, perfect for highly complex processes. As for SCADA? It's a glorified old-school control-room panel whose present-day equivalent is a HMI. Here, the presenter is on the money when it comes to the assimilation of various sources of data and, typically, it sits over a variety of sources such as PLCs, RTUs, and even single and multi-loop panel-mounted controllers (where the latter still may exist). The HMI of a DCS never used to be able to integrate 3rd-party OT sources like dedicated SCADA can. For example, MODBUS transfers would usually involve a dedicated DCS controller card and a communication interface to suit. However, the last time I had anything to do with the ABB 800XA, this SCADA-like HMI was capable of displaying PTX camera feeds, among other things. I'm certain that in the 14 years since, the line between DCS HMI and SCADA has become further blurred in function, with the difference really coming down to the best bang per buck for the predominant BPCS being employed.
Where the data will be stored in this concept if everything goes thought UNS for ML and AI to be able to run the their models. I'm new to this and i see the benefit of UNS but can't understand where the data will live. UNS to me looks like a HUB to pass through data.
Nice presentation. I learned about MES when an organization asked me to apply for an opening as Sql server Database administrator in MES. I am not sure how i can learn more about it. Pls share if you can help.
SCADA was more a term used for line pipe or power line monitoring and management used by utility companies. As a former EMC DCS controls engineer we could hook to anything.
WOW 👏 Tulip good job done !! Please Dont be expressive!! 1) What values in terms of *UNDERSTANDING* the data which was extracted from loads of batch record that the model ran through reading tons of PDF files of batch record? Can you please give OUT a few examples of such outcome you all have witnessed so far ? How many different models do they have or made so far ? No1. Reading the PDF/(paper based? ) or (even digital BRs? in which case do they need a specific format ?) Batch record and bringing out some value ( already asked above if you can describe all / some of those value outcome) No2. Model trained on Technical Specs - have they been generic or for a specific Vertical relating to the plant in which they trained these models. ( Re usability OR START from scratch for every installation) No3. Fault analysis based on (all alarms and events ? from historian ? ) and rectification actions taken to fix those faults. How does the operator describe the fault which model understands either ways even if his description is a bit Skewed by the operator. Requires a lot entry of information by operators , just like giving fault reasons as descriptions by the operators and cleaning that sensibly is another task. No4. *Walkers request* to know the scheduling of various SKUs BUT THAT is not a model , that is information already stored by the production planner in the system !! Its a simple query and nothing to do with LLM !! No5 *Composable MES* app !! INTERESTING need to really dive deep into it , see and learn !! but then which all Industry vertical has their model been trained on to throw out recommendations for ? Surely they would not have covered all possible Industry vertical and processes ( Batch, flow, continuous and custom, etc and then kind of Industry, steel , cement , f&b, pharma, petrochemical .. blah ..blah ) !! At the moment which all have their model been trained on to give good recommendations on!! Calling responses being *Hallucination free* ....NO !!.... may not be the right term; *Verifiable* YES !! but just because you provide reference to its responses, ( point to the teacher 🤣 BLAME IT ON RIO !!) The student STILL must be smart enough to know what is wrong !!!.. Anyways surely a smart move eliminating the cause of hallucinations slowly as the time progresses ( SMART THINKING) 😁
ML was mostly the LLM models they talked about ( A lot of work ( information gathering and feeding ) needs to be done for every installation for these models be really productive) , Nothing was discussed about the prompt engineering ........SCHEDULING was walkers misunderstanding about the function of the model i guess !!
@@ravis2381 little bit supprised about using LLM on scheduling.... optimization problems naturely not fit into ML very well. not sure what kind of complexity level of that planning problem that Tulip can use LLM to help
@@boys7371 I am not from tulip , hope someone from Tulip could answer this , or on behalf of community @Walker could pose this to them and update everyone on the various models they have developed oe @4.0Solutions -> www.youtube.com/@4.0Solutions could help us ! @Tulip www.youtube.com/@TulipInterfaces
Can i redeem my Amazon gift voucher of 60$ one time payment to become a member instead of paying 59 Rs again and again every month via credit card ? Can you explore Amazon payment options also for other things you all provide ? Be innovative😁
Sir, I'm glad that u made video on MES topics.. requesting you to pls make brif video on installation on MES in industry (process, requirement, services, security).. hoping for positive reply.. currently I'm learning MES.. your help would make difference
I watched this video again, and i still don't see Emerson has any DT strategy. 😂 Maybe we need Walker make another video to review Emerson 's DT strategy again.
Data is needed for 4.0 but if you have manual mills and lathes still on your plant floor as many companies still do (JBT is one) the collection of the data is done by the operator manually inputting the production time, parts completed on the shift, setup, scrap, other information into a computer by hand. This is usually done at the end of the shift (bad) and error sneaks into this data so a better way to collect accurate data for these machines is needed, not just run time, downtime, coolant, or lubrication issues but true production data. There will always be error in the data with human input. Legacy companies don't see this as an issue, they still do the same thing they have for years. So If you always do what you have always done, you will always get what you have always got. Their eyes are still closed to 4.0 when it come to data, they think adding a CNC is 4.0 but not collecting the data from it still have operators input of data. Just plan silly to try to get them to understand this as you say the smartest people on the plant floor but they don't make the decisions on how the company runs, someone hundreds or thousands of miles away at a corporate office is. 45 plus years in industry and education with over a half dozen companies working in machining, its a mess out there, legacy companies just don't GET IT.
I think you should change the way you make your videos, cause poeple wants to see and clearly read information on the screen , not your head or your shoulders...this for improvement
Can you point me to the video that you explain about all software for industrial manufacturing, I would like to understand more about what is the Red color of CRM AR/AP SHIPPING WMS meaning , etc , I tried to search all other videos but end up with no specific video for it , Thanks you
The structure of the business and all of the events Single source of truth for all data and information in the business (not system of ownership) The place where current state lives The hub through which smart things in your business communicate The architectural foundation of an Industry 4 and Digital Transformation Initiative (strategy) I think you struggle with the difference between MQTT Broker and UNS. 🙏
it's important to note that the infrastructure is needed in order to train your AI models. As Michael Dell said, without data, you have no AI, and with bad data, you have bad AI. Technically yes, AI Factories is a marketing term, but it's very real how this is happening on the plant floor. Thanks for commenting! Feel free to check out some of our other videos!
Nice Videos! After a long run in plant-floor engineering I recently picked up a job in R&D (both new products and new mfg methods) for a manufacturing organization. The data scene is quite different, and typically straight forward north-stars like OEE have a different meaning here. Do you have any content on how departments that flank manufacturing might get involved in UNS? Thanks!
In a word "defensive". Not willing to acknowledge anything is wrong. He said there's one person (meaning you) against the OPC-F. He agreed with Tom that adoption is king and said they have adoption. Overall a very unimpressive response.
Amazing conversation! Thank you Tom for all of your hard work to start OPC, and transparency on where it could improve. I learned a lot from this conversation, would love to see a pt 2 with Tony Paine
25 plus years in Manufacturing starting in a toolroom to a plant manager and now a consultant in CNC and metals machining and fabrication. I find most manufacturer higher ups don't have a clue as to how the plant floor runs and people on the shop floor don't have much respect for them, just someone in an office pushing paper or messing up the machining production. Then another 25 years teaching manufacturing I find most higher ups don't listen to their shop floor people or education.
We agree, which is why we tell executives constantly that they are not the smartest people in the organization - it’s the people on the plant floor, who do the actual work.
I appreciate the ability to see the behind the scenes of how the OPC sauce was made. I for one would love to see a biography come out showcasing the real unabridged history for each and every industry association - it's helpful information. With that being said, my comment is specifically about the language Tom Burke and many others from automation industry associations state. They all universally talk about “adoption is king, we need to market this better”. To me, this prevailing mindset signifies that industry associations still just don't get it and are likely incapable of ever getting it. It's backwards. You don't sell adoption. Adoption is sold by your users & developers after publicly releasing the asset for developers to meet user needs. OPC and others sell first, worry about meeting actual needs later. This is absolute clownery. Many products exist and are widely known but they are absolute dumpster fires because they don't solve the user's needs. Tom explicitly talks about selling an idea without ever having proven anything to anyone, chiefly by virtue of appealing to big names through status instead of caring about the end user and developers. This is the fundamental difference. True publicized assets - AKA open source technology - prove themselves first by solving real needs and do so by making all associated artifacts are made public. This includes all technical and social artifacts. The mere fact that the specifications are hidden behind paywalls - a childish and petty behavior to say the least - continues to show contempt for transparency and publicized assets. Tom's got a lot of great wisdom, I just hoped to see these aforementioned points covered in much greater detail. It seems to me that symptomatic issues were discussed without addressing the critical, hard hitting topics at hand - namely the several bad actors within the OPC Foundation and its clear political gifts with big consulting companies like Accenture. I was disappointed to not hear Tom discuss the blatantly obvious fact that the OPC Foundation has never been just a “standards” organization but also a lobbying one deeply involved with politics.
@@4.0Solutions As far as calling out bad actors, I like the direct action Giovanni Tarone and Alan Quayle are doing on the telco side. On the app development side, FUTO (although not explicitly labeled as industrial automation) hits directly on the point of ownership and security through open source. Thomas Quieter from the Mobility Independence Foundation is a great champion of true publicized assets and brings a beautiful perspective into why publicized assets are important. My advice to Tom from the OPC Foundation is to come clean and write an autobiography detailing the bad acting of the association. It's time we have serious public whistleblowing on the nefarious acts of prominent industry associations. I'm happy to make introductions to the folks I mentioned earlier.
Loved this podcast. Brought back some memories of the old days. Also remembered when AB/Rockwell bought ICOM. We used their software as well as AB6200 software. ICOM was known back then to put on a great show. I heard stories and got to experience some of them. Thanks, Walker and Tom.
This podcast has laid the foundation for many people and organizations "to do the right things" (as Tom and Walker described) in industrial communication on L2 and will probably be seen as the day the switch was set.
I almost have no words for how much I liked this episode, or how much I learned from it. I think I could listen to stories from Tom for hours. Some of my takeaways: (1) This statement by Tom is pure gold and I will not forget it any time soon: "Success is measured by the level of adoption." That's just how it is, that's real. It's probably why we just can't get rid of Modbus, as outdated as it is. (2) I am guilty of taking it for granted that Classic OPC was invented, but I really did not realize how much work went into getting it done and adopted. Years of work! (3) A lot of cool technical things can be done, but unless it's at a hobbyist level, then the following is essential ... (a) that the solution be made as accessible to implement as possible (b) that it has a real world value proposition ... otherwise you possibly get "so why should I adopt this?" ... question. (4) A possible way forward for the foundation is to go back to the core role of Classic OPC which was process control. Concentrate at getting adoption and the implementation at the embedded level. Thanks for this amazing interview Walker!
Interesting talk, but I think someone needs to go there and fix Tom's ceiling on the right. I think there is a bathroom over there and water is leaking through the floor.
Looking forward to it! Regarding Distribution, at my last event in the US I made a poll on stage to make sure I knew who the audience was, this was the result: 40% vendors , 30% integrators , 15% manufacturers, 15% other
