Artificial Intelligence, always known as AI, along with its sometime companion robots leads the mainstream media hype cycle. It’s going to put everyone out of jobs, destroy civilization as we know it, and probable destroy the planet.
I lived through the Japanese robotic revolution-that-wasn’t in the 80s. Media loved stories about robots taking over and how Japan was going to rule the industrialized world because they had so many. Probing the details told an entirely different story. Japan and the US counted robots differently. What we called simple pick-and-place mechanisms they called robots.
What set Japanese industrial companies apart in those days was not technology. It was management. The Toyota Production Method (aka Lean Manufacturing) turned the manufacturing world on its head.
My take for years based on living in manufacturing and selling and installing automation has been, and still is, that much of this technology actually assisted humans—it performed the dangerous work, removing humans from danger, taking over repetitive tasks that lead to long-term stress related injuries, and performing work humans realistically couldn’t do.
Now for AI. This press release went out the other day, “With AI, humans and machines work smarter and better, together.” So, I was intrigued. How do they define AI and what does it do?
Sensai, an augmented productivity platform for manufacturing operations, recently announced the launch of its pilot program in the United States. Sensai increases throughput and decreases downtime with an AI technology that enables manufacturing operations teams to effectively monitor machinery, accurately diagnose problems before they happen and quickly implement solutions.
The company says it empowers both people and digital transformation using a cloud-based collaboration hub.
“The possibility for momentous change within manufacturing operations through digital transformation is here and now,” said Porfirio Lima, CEO of Sensai. “As an augmented productivity platform, Sensai integrates seamlessly into old or new machinery and instantly maximizes uptime and productivity by harnessing the power of real time data, analytics and predictive AI. Armed with this information, every person involved – from the shop floor to the top floor – has the power to make better and faster decisions to increase productivity. Sensai is a true digital partner for the operations and maintenance team as the manufacturing industry takes the next step in digital transformation.”
By installing a set of non-invasive wireless sensors that interconnect through a smart mesh network of gateways, Sensai collects data through its IIoT Hub, gateways and sensors, and sends it to the cloud or an on-premise location to be processed and secured. Data visualization and collaboration are fostered through user-friendly dashboards, mobile applications and cloud-based connectivity to machinery.
The AI part
Sensai’s differentiator is that it provides a full state of awareness, not only of the current status, but also of the future conditions of the people, assets and processes on the manufacturing floor. Sensai will learn a businesses’ process and systems with coaching from machine operators, process and maintenance engineers. It will then make recommendations based on repeating patterns that were not previously detected. Sensai does this by assessing the team’s experiences and historical data from the knowledge base and cross checking patterns of previous failures against a real-time feed. With this information, Sensai provides recommendations to avoid costly downtime and production shutdowns. Sensai is a true digital peer connecting variables in ways that are not humanly possible to process at the speed required on a today’s modern plant floor.
About the Pilot Program
Participation in Sensai’s pilot program is possible now for interested manufacturers. Already incorporated throughout Metalsa, a leading global manufacturer of automotive structural components, Sensai is set to digitally disrupt the manufacturing industry through AI, including those in automotive, heavy metal and stamping, construction materials, consumer goods and more.
Porfirio Lima, Sensai CEO, answered a number of follow up questions I had. (I hate when I receive press releases with lots of vague benefits and buzz words.)
1. You mention AI, What specifically is meant by AI and how is it used?
Sensai uses many different aspects of Artificial Intelligence. We are specifically focused on machine learning (ML), natural language processing (NLP), deep learning, data science, and predictive analytics. When used together correctly, these tools serve a specific use case allowing us to generate knowledge from the resulting data. We use NLP to enable human and computer interaction helping us derive meaning from human input. We use ML and deep learning to learn from data and create predictive and statistical models. Finally, we use data science and predictive analytics to extract insights from the unstructured data deriving from multiple sources. All of these tools and techniques allow us to cultivate an environment of meaningful data that is coming from people, sensors, programmable logistics controllers (PLCs) and business systems.
2. “Learn processes through operators”—How do you get the input, how do you log it, how does it feed it back?
Our primary sources of data (inputs) are people, sensors, PLCs, and business systems. In the case of people on the shop floor or operators, we created a very intuitive and easy to use interface that they can use on their cellphones or in the Human Machine Interfaces (HMIs) that are installed in their machines, so they can give us feedback about the root causes of failures and machine stoppages. We acquire this data in real-time and utilize complex machine learning algorithms to generate knowledge that people can use in their day-to-day operations. Currently, we offer web and mobile interfaces so that users can quickly consume this knowledge to make decisions. We then store their decisions in our system and correlate it with the existing data allowing us to optimize their decision-making process through time. The more a set of decisions and conditions repeats, the easier for our system is to determine the expected outcome of a given set of data.
3. Pattern? What patterns? How is it derived? Where did the data come from? How is it displayed to managers/engineers?
We create “digital fingerprints” (patterns) with ALL the data we are collecting. These “patterns” allow us to see how indicators look before a failure occurs, enabling us to then predict when another failure will happen. Data comes from the machine operators, the machines or equipment, our sensors, and other systems that have been integrated to Sensai’s IIOT hub.
We trigger alerts to let managers and engineers know that a specific situation is happening. They are then able to review it in their cellphones as a push notification that takes them to a detailed description of the condition in their web browser where they can review more information in depth.
4. What specifically are you looking for from the pilots?
We are not a cumbersome solution, for us is all about staying true about agility and value creation. We look for pilots that can give us four main outcomes:
– Learn more about our customer needs and how to better serve them
– A clear business case that can deliver ROI in less than 6 months after implementation and can begin demonstrating value in less than 3 months.
– A pilot that is easy to scale up and replicate across the organization so we can take the findings from the pilot and capitalize them in a short period of time.
– A pilot that can help Sensai and its customers create a state of suspended disbelief that technology can truly deliver the value that is intended and that can be quickly deployed across the entire organization.
So last week I shared an update on Schneider Electric from the ARC Forum–mostly on cybersecurity. A helpful marketing person guided me to the press release with all the data that updated the software side of the week’s news–specifically asset performance management. For the most part the discussion did not center on product updates but on “increasing momentum surrounding customer adoption”. In other words, Schneider wanted to highlight an area of software not often brought to center stage and show that it is a growth area.
Kim Custeau (I misspelled her name in my last post, I believe–thank you autocorrect), Asset Performance Management Business Lead, shared how investments in the cloud, advanced machine learning, and augmented reality, coupled with new partnerships, have empowered customers.
“Defining and executing an asset performance strategy is a critical component to improving productivity while safeguarding business continuity,” she said. “We have been delivering proven, industry leading asset performance solutions for nearly 30 years, and continue to invest in a long-term strategy to drive innovation in this area. Our focus is to provide real value to our customers by empowering them to maximize return on capital investment and improve profitability. We are proud to see our customer results speak for themselves with significant savings.”
Machine learning and prescriptive analytics:
- Duke Energy prevented an estimated $35 million cost from early warning detection of a steam turbine problem
- Ascend Performance Materials now responds faster to alerts saving an estimated $2 million through avoided plant shutdowns
- BASF is implementing AR to improve asset performance, reliability, and utilization while increasing production efficiency and safety because technicians leverage an augmented digital representation of the asset.
Cloud and Hybrid Deployment:
- WaterForce partnered with Schneider Electric to develop and IIoT remote monitoring and control system in the cloud that allows farmers to operate irrigation pivots with greater agility, efficiency, and sustainability.
- MaxGrip and Schneider Electric announced a partnership to expand APM consulting and add Risk-based Maintenance capabilities. The APM Assessment is a first step for industrial companies to evaluate asset reliability and digital transformation strategy.
- Schneider Electric and Accenture completed development of a Digital Services Factory to rapidly build and scale new predictive maintenance, asset monitoring, and energy optimization offerings. As a result, a large food and beverage company saved over $1 million in maintenance costs
Cybersecurity, digitalization, and asset performance management headlined the various press events with Schneider Electric at the recent ARC Forum. I took notes from Kim Cousteau’s presentation on APM at the main press conference and expected a follow up press release for details. I have not received one yet.
Remember the “reverse acquisition” of Aveva where Schneider Electric placed all of its software divisions into Aveva and then took a 60% share in the company? The deal is about to close. Schneider spokespeople assured me that digitalization is proceeding apace with the leveraging of Aveva design through construction applications into operations and maintenance applications—Schneider’s strong suit. This, on paper, brings the company into the competitive marketplace with Siemens and its UGS acquisition of several years ago. This is an interesting area to watch.
Schneider called a special press event, with lunch, to talk specifically about cybersecurity. This response to an incident in which the company’s Triconex safety system earned some publicity—but not always accurately portrayed. The incident was a cyber attack that caused a situation that the safety system caught and initiated a safe shut down.
However, the event caused renewed concern for cyber defense. ARC Vice President, Larry O’Brien, said, “This is a wake up call for people to follow existing security standards.” Gary Freburger, who heads that division of Schneider, said, “It’s everybody’s job.”
We received this official statement from Peter Martin, vice president of business innovation and marketing, Schneider Electric
At Schneider Electric, we heartily encourage all collaborative efforts to strengthen cybersecurity. The growing problem of cybersecurity is not specific to any single company, institution or country. Rather, it’s a threat to business and public safety that can only be addressed and resolved when suppliers, customers, integrators, developers, standards bodies and government agencies work together. This collaboration starts with common standards, agreed-upon rules, appropriate funding and active cooperation. It extends beyond national borders and transcends competitive interests.
Schneider Electric continues to work diligently with our customers, partners, developers and industry peers to make the shift from reactive to proactive cybersecurity management through compliance with evolving industry standards, agreement that cybersecurity is a journey not a destination, and a commitment to standing together in the face of cyber threats.
Today, we commend the signatories to the “Charter of Trust.” It’s another important step toward ensuring that the promise of digital transformation and automation will prevail over the threat of cyberterrorism.
Regarding APM, Kim Cousteau discussed a new release of Avantis that expanded machine learning from the power industry to oil & gas. For maintenance, it incorporates a team system for operator rounds and improved workflow. It incorporates augmented reality and virtual reality (AR/VR) “because workers are so new and need help to get up to speed. Look for updated analytics to aid in catching anomalies ahead of failure. She cited a customer who has been tracking savings from this feature alone and is up to $65 million.
The product Rockwell Automation executives most wanted to talk to me about at the last Automation Fair event was its new analytics platform.
Immediately following the Rockwell event was Thanksgiving, then a trip to Madrid for a Hewlett Packard Enterprise event followed by catching up and Christmas. But I grabbed moments to contemplate the “Project Scio Edge Analytics Platform” (see image) and tried to place it in a context amongst all the platforms I saw this year. Which were many.
Executives including SVP and CTO Sujeet Chand and VP of Information Software John Genovesi were enthused over the new product. I wrote about it here.
I liked much of what I heard. There were many overtures to open connectivity that I have not heard at a Rockwell event—maybe ever. I even got an hour to discuss OPC UA and how Rockwell now intends to implement it. The demo during media days was also powerful.
I drew a mind map and exported an outline. Here is the list of positive things.
- Developed analytics from acquihire
- Good UX
- Open connectivity including OPC UA
- Should provide customers with insights into control systems and machine performance
However, I’m left with some questions—some of the same ones I often feel about Rockwell Automation. Check out the architecture diagram. It stops with machine level. I always expect to see more, but Rockwell always stops at the machine. Perhaps GE and Siemens have overreached with Predix and Mindsphere (and Schneider Electric with EcoStruxure?), so Rockwell stays closer to its roots on the plant floor? Is it more profitable and manageable that way?
I don’t know the answers. But I’m left thinking that with the rise of platforms [see for example Platform Revolution by Geoffrey G. Parker, Marshall W. Van Alstyne, and Sangeet Paul Choudary] and open ecosystems, Rockwell seems to have a much smaller vision. It talks of “Connected Enterprise”, but in the end I don’t see a lot of “enterprise” in the offerings.
- Is it platform or a piece of the Rockwell software stack that stops short of plantwide views?
- Is it anything that others (SIs and users?) can add to?
- Is there more coming?
- Is there a way to integrate supply chain and customer chain?
- Seems a natural to integrate with an asset management application–which Rockwell does not have.
I think they’ve done well for what they evidently set out to do. I also think there remains more to do to help customers leverage the Internet of Things and Digital Transformation. Interesting to see what next November brings.
Efficiency—running a plant with maximum planned output, minimum waste, and best use of people and assets. A friend has refreshed and refocused his magazine on those concepts.
Efficient Plant is the new generation of Maintenance Technology magazine. It includes the best of what they have done for years and upgraded to reflect advances in control, automation, Internet of Things, software, and strategies.
Oh, and after a three-year hiatus, I am again writing a column every month.
Check it out—and let us know what you think.
GE has a new CEO coming soon. Jeff Imelt led the industrial push that led to Predix and Industrial Internet of Things, services based upon data, predictive maintenance. He spurred development of GE Digital and the transformation into a software company (check out the TV ads).
The company has announced some extensions to Predix. But we need to wonder where the new CEO will take the company. One software entrepreneur I know unleashed on the company in a LinkedIn post hoping that the new guy would trash Predix and build “something better.” We’ll have to wait and see, of course.
First a little context for one of the announcements.
The facility electrical engineer and I were speculating on an idea of linking measurement of electricity usage at perhaps the bus level for different areas of the plant with machine performance. Perhaps he could detect a machine problem through electrical changes. That was somewhere around 1993.
I quoted something, but we never did it.
Here are the high points of the announcements:
- GE Digital announces integration of ServiceMax field service management solution and Asset Performance Management portfolio to transform service operations, reduce cost and eliminate unplanned downtime
- GE Ventures launches Avitas Systems, a new venture that will transform inspection services with advanced robotics, data analytics and artificial intelligence
- GE Power releases Predix-powered ‘Digital Utility’ to connect real-time machine and operations data with energy trading to drive more profitable utilities businesses
The asset optimization organizations within a plant have a variety of new tools to take them beyond maintenance into an enhanced role. Growth of the Internet of Things and analytics capabilities especially leading into predictive and eventually prescriptive strategies are the keys to the future.