(Disclaimer: If you have already deployed a robust video surveillance system for your plant or warehouse operations and you are using it aggressively for process improvement … “never mind” … read this in case it contains some useful ideas, but it is not aimed at you).
As previous posts have argued, I believe that manufacturing and logistics operations are grossly under-utilizing video technology in their management and operations. I believe there are many easy wins if you apply standard video surveillance technology to current manufacturing and logistics processes. However, as I watch the growing momentum behind Industry 4.0 (and Industrial Internet Consortium, and …), I am convinced that video may be even more critical to success in those ventures
Implementation Challenges for Industry 4.0
I was around in the 1980s when GM bought EDS and introduced early automation efforts like Saginaw Gear and CIM and MAP 1/2/3. I remember the endless discussions about “islands of automation”. I’m confident that the data connectivity problems of that era are well and truly solved. But I also remember the angst that occurred when engineers tried to integrate new technology into legacy production systems. It wasn’t just the data or technology interfaces … there were serious personnel and cultural issues. The computer geeks didn’t communicate well with the assembly plant engineers … or maybe the other way round. It mostly wasn’t personal animosity. It was the fact that neither group understood the other’s tools and technology … to the point they seemed to speak different languages.
Fast forward and Industry 4.0 is proposing tectonic shifts in manufacturing processes and technology. The shift is so big that virtually all consulting documents share two expectations:
Industry 4.0 will demand a new workforce:
The skills needed for Industry 4.0 will be very different from the ones that are successful in conventional manufacturing. Conventional manufacturing is designed to break work down into well-specified processes so that modestly educated and trained workers can execute them consistently and correctly. Implicit in this model is the assumption that most work is managed visually. With Industry 4.0, most of the work will happen inside machines or in algorithms that use data and are hidden from the human eye. Everyone from design engineers to managers, to front-line workers must be able to read and understand the data flowing through the system. That will be the new definition of “sight”.
- “The critical Industry 4.0 jobs—for example, data managers and scientists, software developers, and analytics experts—require skills that are fundamentally different from those that most industrial workers possess today. ” – BCG – Inside-OPS-Jul-2016.pdf, p6
“Lack of skills or competencies in the company’s workforce is also the biggest challenge survey respondents see when it comes to making use of data analytics.” – PwC – industry-4.0-building-your-digital-enterprise-april-2016.pdf, p17
- “In the process of trying to integrate IT and OT through the use of Industry 4.0 practices at the organization level, companies often face a shortage of talent to plan, execute, and maintain new systems. The number of engineers trained in handling unstructured data and big data tools—crucial for the type and scale of data generated by connected systems—is gradually increasing, but still falls far short of anticipated demand.” Deloitte University Press – DUP_2898_Industry4.0ManufacturingEcosystems.pdf, p17
Industry 4.0 will be implemented incrementally
Designing new equipment, integrating new systems and training or retraining a new workforce will take time. Probably a fairly long time. Also progress will be made at different rates for different types of manufacturing. Some products may be ready for early Industry 4.0 now. Others will require significant further development. It is highly unlikely that any rational company will leap off a cliff into a wholesale adoption of Industry 4.0. There are too many personnel and management changes to understand and absorb.
“The shift to Industry 4.0 will not happen suddenly and will not necessarily stick to ideal and typical change management models.” – KPMG – factory-future-industry-4.0.pdf , p31
- “Replacement of existing tools with machines during the third industrial revolution (automation) is estimated at up to 90 percent. Under Industry 4.0, we believe that the main requirement will be upgrading existing equipment, mainly in the dimensions of sensors and connectivity.” – McKinsey – mck_industry_40_report.pdf, p14
- “Unlike prior industrial revolutions, Industry 4.0 is not about replacing the existing assets with new ones but about mastering the managerial challenges …” – McKinsey – mck_industry_40_report.pdf, p15
Industry 4.0 won’t suddenly happen like this:
It won’t even happen smoothly like this:
It’s most likely to happen in a complex patchwork quilt like this:
During this transition, key production assets will continue to operate on the old labor-based, eyeball and paper management systems. Gradually, they will switch to the new digital, data-driven system. How will they be coordinated and managed through this transition?
Why A Comprehensive Video Surveillance Strategy Could Help Management Immensely
First, what do I mean by “video for management”? After all, specialized video sources like machine vision are already intrinsic to Industry 4.0.
I mean the high end video security systems that are deployed in offices, factories, other buildings, on city roads, in railway yards, at warehouses, at shipping terminals … the standard security systems that are sold by the thousands every week. These systems are designed to monitor sections of facilities and people-scaled activities. They are getting better and cheaper every day … and their cameras can see clearer, farther and closer that ever before … but they are still intended mainly as a management tool. Today, they are used to manage security issues … but they could be used to manage operations. They are the systems with user interfaces like this:
A system like this would be deployed with different types of cameras watching different parts of the operation … for different reasons.
Some high-resolution cameras would be mounted to take a broad view. 4K resolution cameras will capture a lot of detail, even from a distance. You can use digital zoom to see something specific … and do it at any later date. The rest of the cameras would be deployed to get close ups of points of interest. The network video recorder collects the video feeds from all of the cameras and stores them on hard drives … possibly for weeks before the storage must be overwritten.
Recommended Management Uses for Video Surveillance
As a previous blog article briefly mentioned, video security surveillance are typically used in two modes: real-time monitoring and forensic record-keeping. In the manufacturing context, especially Industry 4.0, I would add a third use: remote collaboration.
NOT for real-time Monitoring
I would mostly ignore the idea of real-time monitoring for factory management. It’s labor-intensive and won’t find much unless someone knowledgeable is watching the monitors. Why place an expert in a back room? Wouldn’t they be better on the floor helping to organize things, or in an office planning near-future production? Real-time monitoring also carries the stigma of “big brother watching” that won’t help morale. I know a smaller factory that put a couple of surveillance cameras on the floor so the back office can locate the manager when he wanders the floor. But no one watches the feeds continuously.
Comprehensive video record-keeping is an huge opportunity
To me, this is a manager’s dream. Effectively, it is a “time machine” that can return you to the moment when that now-evident quality, safety or production issue actually happened. How many factories maintain SPC charts to watch for process changes? Wouldn’t they love to see what was happening nearby when that machine tool started to behave erratically? Wouldn’t they want to watch that specific part (the one that showed as defective at final inspection) throughout its journey in the production process? Was it handled differently? Where was it stored?
Police use this technique countless times every day to solve crimes. They review the security footage to see if it contains a clue. Why is a production anomaly any less of a “crime” to a business that is trying to survive?
Even if you aren’t hunting a problem or a defect, wouldn’t it be useful to have video of the process as it actually operates. Industrial engineers often bring cameras to the process when they are trying to do safety or efficiency studies. Why bring cameras? If you have a robust surveillance system, the cameras may already be in place and the video may already be on file. Just find it and go from there.
No one likes to write SOPs and work instructions. Everyone does it because it’s the heart and soul of standard work and it forms the foundation for training. But … imagine if you could just go and experiment with different workflows and work methods and machine settings until you found something better. If you had a robust video surveillance system, you wouldn’t have to try to remember what you did. You would pull the video for that period and area and write the SOP as you watched what you actually did … that really worked. Faster, easier, and more accurate. You could possibly delegate it to an assistant … or an external contractor.
This brings me back to Industry 4.0 and the patchwork quilt. I expect the technology that is now being designed and deployed will contain amazing data collection and diagnostic systems. They will “see” facets of the production process that no camera or eye could ever see. But will they be able to look around and see what is happening nearby? Will they see that the bar code labels on the shipment are a bit off-center? Will they watch the contractors bump them slightly with the lift truck as they install the machine next door? Will they be watching the flex in the extension arm that the designer never thought would be a problem?
If I’m right, the times of greatest concern will happen when new systems are installed, tested and commissioned … and at the boundaries where the new Industry 4.0 facilities interact with legacy operations. The new systems will be installed in facilities that don’t emphasize old-school visual management. I worry that the people who know how to run Industry 4.0 systems may not be as aware of their physical surroundings as the eagle-eyed supervisors in a conventional plant. I’m sure they will be very sensitive to anomalies in the data, but will they really be as sensitive to clues in their physical environment? I have worked with some extraordinary controls and systems engineers. I know some who see everything. But they are rare. If it were me, I would want to be able to tap into a robust video surveillance system … if nothing else, as insurance.
Remote collaboration may be the biggest win of all
Look again at the last (patchwork quilt) video. One of the signatures of Industry 4.0 is that data will tie together the entire supply and value chain. Things will happen in upstream suppliers that are directly tied to downstream assemblers and possibly impact end-of-distribution retailing. When everyone is so interdependent, wouldn’t it be great if they could collaborate over a live (or recent) video of the process in question? I expect that for large companies, Industry 4.0 will be deployed across an entire value chain for a given product … including internationally … before it is applied to all of the different products in a given plant.
If I am right, the same technical journey will be happening in plants around the world … around the same time. Wouldn’t it be great if the engineers in the next installation could have direct reference to the surveillance video of the plant that immediately preceded them with the same process and equipment. Wouldn’t it be great if the engineers at the first plant could flag a process design issue and share the video with their successors … before their successors fell into that trap themselves?
What Will It Take to Start?
Building these systems should be dead simple and easily affordable in current operations, let alone compared to the cost of Industry 4.0. Just call your current security systems vendor and have them give you a quote to extend their systems to cover operations inside your walls. There are a lot of common objections, but I don’t see any real risks and I have tried to lay out the case for the benefits. Once you have a capture system installed and running, you can take stop and take a breath. The video will be flowing in. Now all you have to do is figure out what to do with it?
Where do you go from there?
Once the video stream starts to flow, I have two suggestions. First, start to look at the video and see what you find. Maybe move a few cameras around. Try looking from a distance and close up inside machines. Try looking for unusual events … both good and bad. Try watching a video feed really carefully to count or measure part of your process performance.
Second (or maybe first), contact DeeperPoint and let us help you start this journey!