If you look at the major industrial sensor manufacturers (e.g., IFM), many of them increasingly don't manufacturing products without IO-Link. They've long had the technology in their sensors to do some basic logic and configure the output (e.g., 24VDC, 4-20mA, 0-10VDC). Many people don't realize they likely already have IO-Link sensors in their plant. The adoption of IO-Link has definitely picked up in recent years. I believe the major inhibitors are legacy technology and people. A lot of plants still have old PLCs and/or have adopted ethernet fieldbuses. The other factor is that many people do not care about process monitoring. They will equip machines with as few sensors as possible and will use digital outputs. If you ask a machine builder for variable process data (e.g., IO-Link, 4-20mA analog), they will find every excuse not to. They will make sure it breaks your budget and delays delivery. It can get exhausting explaining to people that there is nothing predict when the only signals from the field are 'On' or 'Off'. No apps or analytics are going to inform anyone how 'On' or 'Off' something is if it's only 'On' -or- 'Off'.
The raspberry Pi isn't rated for Ex/IS environments. Same for almost all other IIOT "smart" sensors. The "someone" who wants to build this wonder sensor is going to have spend big $$ to certify the damned things. In non-critical environments things will change relatively fast but in critical safety/ hazardous environments it will take some time before plant safety / process control boards are willing to bet lives and dollars on it. Great video nonetheless.
Industrial automation and control may look something like this in a few decades, but you have to temper your wild enthusiasm with realism. No industrial plant is going to rip and replace all of their production sensors and control devices just so it can report non-critical data like it's temperature or firmware revision to a unified database. The cost is vastly prohibitive and the benefit is only theoretical. A new sensor may only cost $500 but 1000 new sensors? And the time it takes to install, test and commission them? It's more likely that new installations will incorporate these components on an as-need basis. I've worked on several greenfield facilities in the past 3 years and I've seen zero IIOT or 4.0 or such ideas implemented. Some companies are developing data lakes, which is a good thing, but even that has been laden with bad design and an almost total lack of comprehension of the value of time series data. Think about this in terms of Industry 3.0. Believe it or not there are still, STILL, some facilities running equipment with local electro-pneumatic controllers or isolated PLC's. And there are perhaps thousands of vintage PLC's and DCS's running in the US alone. Many of these were commissioned before I was working in industry, over 30 years ago. That represents billions of dollars in infrastructure. I wish we could get more of them to rip and replace!
Articulated in a very practical way....Fantastic, Great Job... What I observe in the field, there are no takers of IO-Link technology, Automation Vendors are pushing Software piece only like Thingworx or Mindsphere.
I really like the video and the point you make. But I wouldn't use the Pi as the example of something to add in an industry environment. However, I understand it helped make the point, so good!
Most of the imbedded data in devices that you reference has been available for decades via HART and other carrier wave protocols but plant personnel have never seen the value in having that information. I have personally struggled to convince management that this data might be of use for years. I think it's because for the vast majority of sensors & control elements, most of this data has little value beyond the concepts of predictive/preventative maintenance. Sure, Tesla is probably highly invested in this sort of data, but is a glue manufacturer, a paper mill or a steel plant really interested in it? Not that I've seen. Go into any heavy industry plant and you can easily find scores of devices that are woefully out of date, many that don't work at all and lots that are configured incorrectly. And yet the plant still runs just fine. Convince management to fix or replace all those devices that barely work first, then let's talk about connectivity.
I am seriously skeptical about the stability of those devices when they are made IOT ready, by the introduction of IOT the circuitry of each individual field devices will become complex and there by the chances of failure can become high.
biggest drawback for a large plant having IIOT is being able to process all that data especially when you get into the 1000s of tags in a HMI system. having to pull this information & trend the devices can slow down the system alot.
This is the primary reason the minimum technical requirements for your IIoT protocol are edge driven, report by exception, lightweight, open architecture.
Q: Will IIOT sensors still have wired connections to the PLCs? Or all sensor data will go to the Internet and then collected by the PLCs from the Internet (I am asking because of data speed transmission).
![[LIVE] : ONE ลุมพินี 84 วันนี้!! คู่เอก "ก้องศึก vs เมืองไทย"](http://i.ytimg.com/vi/3kuzsAIDf1c/mqdefault.jpg)
Another clear & easy to understand video, thank you
Glad it was helpful!
Walker you are a truly fantastic educator!
Thank you!
If you look at the major industrial sensor manufacturers (e.g., IFM), many of them increasingly don't manufacturing products without IO-Link. They've long had the technology in their sensors to do some basic logic and configure the output (e.g., 24VDC, 4-20mA, 0-10VDC). Many people don't realize they likely already have IO-Link sensors in their plant.
The adoption of IO-Link has definitely picked up in recent years. I believe the major inhibitors are legacy technology and people. A lot of plants still have old PLCs and/or have adopted ethernet fieldbuses. The other factor is that many people do not care about process monitoring. They will equip machines with as few sensors as possible and will use digital outputs. If you ask a machine builder for variable process data (e.g., IO-Link, 4-20mA analog), they will find every excuse not to. They will make sure it breaks your budget and delays delivery.
It can get exhausting explaining to people that there is nothing predict when the only signals from the field are 'On' or 'Off'. No apps or analytics are going to inform anyone how 'On' or 'Off' something is if it's only 'On' -or- 'Off'.
Thank you for sharing about IO link! Have you already joined discord?
The raspberry Pi isn't rated for Ex/IS environments. Same for almost all other IIOT "smart" sensors. The "someone" who wants to build this wonder sensor is going to have spend big $$ to certify the damned things. In non-critical environments things will change relatively fast but in critical safety/ hazardous environments it will take some time before plant safety / process control boards are willing to bet lives and dollars on it. Great video nonetheless.
Industrial automation and control may look something like this in a few decades, but you have to temper your wild enthusiasm with realism. No industrial plant is going to rip and replace all of their production sensors and control devices just so it can report non-critical data like it's temperature or firmware revision to a unified database. The cost is vastly prohibitive and the benefit is only theoretical. A new sensor may only cost $500 but 1000 new sensors? And the time it takes to install, test and commission them? It's more likely that new installations will incorporate these components on an as-need basis. I've worked on several greenfield facilities in the past 3 years and I've seen zero IIOT or 4.0 or such ideas implemented. Some companies are developing data lakes, which is a good thing, but even that has been laden with bad design and an almost total lack of comprehension of the value of time series data.
Think about this in terms of Industry 3.0. Believe it or not there are still, STILL, some facilities running equipment with local electro-pneumatic controllers or isolated PLC's. And there are perhaps thousands of vintage PLC's and DCS's running in the US alone. Many of these were commissioned before I was working in industry, over 30 years ago. That represents billions of dollars in infrastructure. I wish we could get more of them to rip and replace!
Thank you for your comments Mark. Check out this video for our responses th-cam.com/video/GruN5S18-qI/w-d-xo.html
Drops mic! Well done fellas!!
I love that part 2
As if these ideas are new...
Articulated in a very practical way....Fantastic, Great Job...
What I observe in the field, there are no takers of IO-Link technology, Automation Vendors are pushing Software piece only like Thingworx or Mindsphere.
Thanks for sharing. There was actually someone on Zack's podcast who mentioned IO link. Rob Sawicki.
Thank you, It is very useful
You are welcome
Another great video, what is your view on io-link? Sensors with intelligence are becoming more valuable, is io-link a step in the correct direction?
Yes! Companies like Pepperl and Fuchs are creating IO link devices that publish directly to MQTT!
Amazing!
Thank you!
Great Job Walker! My question is: What companies or Brands for instrumentation and sensors are working with IIoT?
New brands every day! IFM has some good ones.
Banner is another one to look at.
Right On Point!
Thank you!
I really like the video and the point you make. But I wouldn't use the Pi as the example of something to add in an industry environment. However, I understand it helped make the point, so good!
Thank you Rafael!
Hats off for video🤗❤🇮🇳, Btw i have a que, How long PLC will exist ?? OR Are they gonna die after a decade ??
Great question! I personally believe some form of PLC or something similar will exist at the edge for safety critical processes.
Most of the imbedded data in devices that you reference has been available for decades via HART and other carrier wave protocols but plant personnel have never seen the value in having that information. I have personally struggled to convince management that this data might be of use for years. I think it's because for the vast majority of sensors & control elements, most of this data has little value beyond the concepts of predictive/preventative maintenance. Sure, Tesla is probably highly invested in this sort of data, but is a glue manufacturer, a paper mill or a steel plant really interested in it? Not that I've seen. Go into any heavy industry plant and you can easily find scores of devices that are woefully out of date, many that don't work at all and lots that are configured incorrectly. And yet the plant still runs just fine. Convince management to fix or replace all those devices that barely work first, then let's talk about connectivity.
Thanks for the question Mark! We responded in this video here: th-cam.com/video/GruN5S18-qI/w-d-xo.html
I am seriously skeptical about the stability of those devices when they are made IOT ready, by the introduction of IOT the circuitry of each individual field devices will become complex and there by the chances of failure can become high.
We will do a video on this.
Thanks for the video
Of course!
biggest drawback for a large plant having IIOT is being able to process all that data especially when you get into the 1000s of tags in a HMI system. having to pull this information & trend the devices can slow down the system alot.
This is the primary reason the minimum technical requirements for your IIoT protocol are edge driven, report by exception, lightweight, open architecture.
Q: Will IIOT sensors still have wired connections to the PLCs? Or all sensor data will go to the Internet and then collected by the PLCs from the Internet (I am asking because of data speed transmission).
Most will publish wirelessly to the cloud. By percentage. Many will be wired into a PLC because of time and latency requirements by application.
So should I do my masters in data analytics or mechatronics ??
Is up to you. Data analytics isn't going away anytime soon!
@@4.0Solutions is mechatronic instead going away?
Well, I think you should sit back, take a cup of tea/coffee and ask yourself that question every day until you find the answer.