For decades, manufacturing and production operations created masses of data from sensors, control strategies, diagnostics, procedures and other applications used to safely control the process. Operational analytics have continued to advance to turn this mass of data into actionable information.
Emerson’s Manasi Menon joins me in this 23-minute operational analytics podcast to discuss how organizations are putting these analytics into action to help improve performance in safety, reliability, energy & emissions, and production.
Jim: Hi, everyone. This is Jim Cahill, and welcome to another Emerson Automation Experts podcast. Today, I’m joined by Manasi Menon. Manasi has a Bachelor of Technology degree in electronics and instrumentation and also an MBA. She’s worked for several automation suppliers and has been with us at Emerson for the past eight years. Manasi currently leads our machine learning and analytics marketing efforts. Welcome, Manasi.
Manasi: Thank you, Jim.
Jim: Well, let’s get started. I gave a little bit of some of your background, but can you give us a little bit more and your path through working with the automation suppliers and Emerson up to your current role with our analytics?
Manasi: Sure, Jim. As you mentioned in the introduction, I am an electronics engineer who’s worked very closely with some of our customers in this process automation industry. There is a lot of interest and there is a lot of talking around analytics today, but our customers have always been using analytics as an opportunity to improve their processes. So, my introduction to analytics was back in the days as a process engineer where I was working for end users who were using conventional advanced process control systems to improve their process, whether it is optimizing their process or monitoring the process. So, right from that time frame, after that, doing my MBA, I ended up in a particular position where I was working as an analyst monitoring the support functions of the organization. So, with that process engineering background and an interest and passion for analytics, I thought this role would be really good where I could contribute to as well as learn from, especially articulating the value add of our analytics in our industry.
Jim: Well, that’s interesting that so many more of our customers are looking at digital transformation and doing something with this wealth of data that they’ve always been collecting. So, I imagine there’s a lot of people on our customers’ side coming into this area of having more responsibilities for analytics. So, I guess with all that, for someone new into the consideration process for additional analytics as part of their automation architecture, what are some ways that they can get started?
Manasi: So, as part of my role, I talk to a lot of customers, some of them who have advanced quite a bit in the stage of deploying analytics in their organization and some of them who are still new. One of the most important things that I have seen works really well when thinking about deploying analytics in their organization is to have a clear goal, a strategy or a clear business goal in mind as to why they want to deploy analytics. Of course, what is also important is once you’ve identified the goal, understand what kind of sponsorship do they have within their organization, to adopt that goal and replicate it. So, that is first step. Then the second step is to identify a clear use case. And it depends on where you are in the organization. It’s very important to collaborate with folks within your organization. And when I say collaborate, what that means is work with plant managers, with the operators, with the reliability engineers because these are the people who are going through those problems, which that analytical solution can resolve.
So, once you talk to them, interview them, you would understand what use case you would like to deploy the analytics to. Then mostly these use cases are very easily replicable or customizable or easily you could deploy it in multiple facets of your organization. So, that’s one consideration to choose that use case. Then when you do that, then you would think about, okay, what does the technology infrastructure that I need to make this possible? Whether it is data, whether it is network or the software application itself that you would use. That’s the typical path that we have seen a successful organization, or a successful customer would do to deploy analytics.
Jim: Can you tell us a little bit more about the different types of analytics that can be applied?
Manasi: So Jim, analytics is extremely broad. Per Clean Energy Smart Manufacturing Institute, there are about over 600 or so analytical vendors out there that offer some types of analytics, whether it is artificial intelligence or machine learning or optical character recognition. So, there are just so much, so much out there. Now, within that, we feel that that space can be divided into two key areas. One is business analytics, focusing on HR, supply chain, finance, and CRM and one is operational analytics, which is I talked about is what Emerson is focusing on, which is the day-to-day operations of the plant. Now, within the operational analytics, there are again two different kinds of analytics depending on the problem that you’re trying to solve. So, if you’re trying to solve the problem of a very well-known asset or a very well-known use case where there is enough subject matter expertise around it, then it’s called principles-driven analytics. If you’re trying to solve a problem that you do not know why it is happening and it may be happening because of the way multiple parameters or multiple things are correlating and interacting with themselves, then we call it data-driven analytics. So, within this operational realm itself, there are two different kinds of analytics.
Now, Emerson has a very, very, very strong background and history of analytics. We’ve had a lot of analytical offerings from 40 years ago, like with AMS Device Manager where we have analytics embedded in our diagnostics devices, to analytics which is embedded within our automation system such as DeltaV, whether it is machinery-specific analytics such as AMS Machinery Manager, or that principles-driven analytics that I talk about, such as Plantweb Insight, which is helping our customers to get started really easy with some of their asset-specific analytics, which is monitoring their health and performance of the asset, to data-driven analytics where you’re trying to solve a problem that you do not know why it’s happening. So, we have Emerson’s KNet that offers advanced analytics techniques in combination with some of these principles-driven analytics to give you a more holistic approach as to why a problem is happening. So, Emerson has a very strong and a variety of offerings to cater to each of these needs. And of course, added to that, which I did not mention and I want to make sure is, is there as part of this whole solution is our AMS Asset Monitor, which is focusing on edge analytics. So, those are the different types of analytics and our different offerings that we have.
Jim: So, what are some examples where other manufacturers and producers are applying analytics and receiving some value from it?
Manasi: So, Emerson’s analytics offering is focusing on what we call operational analytics, which means focusing on the day-to-day operations of the plants itself. So, some of the use cases that we have seen our customers deploying analytics to is in this realm of operational analytics, specifically in the area of process, energy, reliability, and safety. So, let me try and give you an example of each. So, when it comes to process, it is using analytics to monitor your process, whether you’re monitoring the quality of the product and ensuring if it is deviating from a standard spec. So, you’re doing quality control, or you are trying to optimize your process where the analytical application is monitoring the daily operations, change in the operating modes, change in the way the plant is functioning, and recommending a certain set point or recommending a certain way that that plant should operate. So, that’s one area.
The other area where we have seen analytics being applied to is in the case of energy. When I say that, what I mean is in this day and age, every organization, every company has a certain target that they want to meet in terms of energy consumption. So, using analytics to monitor, to calculate, to estimate, and also ensure that the energy consumption of your plant is not deviating from a certain target that you are requiring to meet. So, that’s one example where the analytics application has been deployed by the customer. And the other realm that I talked about is reliability. So, monitoring the health of the asset. And it’s not just monitoring the health of the asset because that’s been done for quite a few years using various pieces of technology, but there is predicting whether an asset is going to go bad, and also using analytics to prescribe as to what should be done when that asset is going bad, so monitoring the health of the asset and monitoring the performance of the asset.
And we’ve actually seen our customers use analytics to start with few and then expand that to a fleet of assets. They tend to do monitoring of the entire fleet in terms of a reliability perspective. And of course, the final is the safety, which is extremely key for any plant in our automation industry. So, ensuring the use of analytics to predict and also to make sure that they can execute a certain procedure in a very safe manner using analytics. So, those are some of the examples where we have seen customers are deploying and getting the value from.
Jim: So, definitely no shortage of opportunities when you look at production, you look at reliability, you look at safety, you look at… What was the one I’m missing?
Jim: Energy. That’s right. Energy efficiency. So, all kinds of examples. So, if a person picks one of those areas and says this is the part that we’re gonna focus on and try to get some value, some early successes and build from there, what are some important considerations in developing a plan and moving forward?
Manasi: So, let’s say they followed all those steps and they have identified a use case. What they also should do is to have a clear business metric that would warrant the success of that use case. So, have a certain ROI in mind. After that, lay the foundation for the technology. For analytics to work well, the underlying infrastructure of data is extremely important. So, work towards developing that, whether it is data governance or whether it is data storage or whether it is establishing the network architecture to where the use case applies to. So, investing in that. After that, we know you would typically work with a certain vendor to deploy that use case for analytics. And let’s say you’ve done that and you have established a clear business metric or you’ve seen the value add that that analytical solution has provided, the next step that we have seen our customers do is establish a center of excellence, of experts from different parts of the organization who are catering to each of these use cases.
What we have also seen is them partnering with vendors such as us as part of this center of excellence for analytics and then scaling that up to different levels and different…whether they want to move up to a plant level or to multiple plants at an enterprise…or at an enterprise level. So, continue to develop, deploy, find use case, elaborate, and then expand on that. And then finally, also, educate. Each of these…the end users who are using the results of the analytics themselves need to understand how it works and which insights they should tap into. So, continue to circulate that information, which the analytical application is spitting out as in the form of reports, in the form of information among different users and use that in the daily processes.
Jim: Okay. And early on you had mentioned the importance of it being cross-functional, not just one group leading the effort to be successful with it. So, which roles within a company typically get involved in the decisions that need to be made?
Manasi: What we have seen is it is usually the corporate that has an initiative and is taking the ownership in terms of sponsorship and implementing it. When I say corporate, I mean the office of the CIO or the CEO or the COO. But, what we’ve also seen is the individual executives of the plant, whether it is the GM or the VP of the process optimization, or the reliability or the safety or the utility all working together very closely with the vendors themselves and influencing that decision back to the corporate and determining which vendor to go with. So, it’s extremely important to work along with them as well who would ultimately be your champions to get that going within an organization.
Jim: Well, that makes a lot of sense, then bringing people together and getting those various viewpoints to be successful. So, are there any specific examples you can share, of course, without naming names, of companies that have added analytics to one of those areas and seen some quantified results?
Manasi: There is one particular such use case that comes to mind. It is a chemical company, I’m not naming it, but essentially, they followed the same process that I just talked about. They had a corporate initiative to deploy analytics within their organization. What they did was they established a center of excellence which comprised of certain stakeholder from different parts of the organization to come together and brainstorm as to where they would want to apply analytics to. So, this group of body evaluated I think over 50 use cases and then they determined, “Let’s start with the two use cases. And then once we feel that we have achieved that ROI out of it, then we would expand to more.” So, they decided to focus on A, reliability, and B, it was process monitoring. There is a very important reason reliability because it was a repeat concern for them, and it was causing unwanted shutdowns. I think this particular use case caused at least two shutdowns for them, which was unwanted. Process monitoring because the particular scenario that they were dealing with, it was costing them a lot in terms of the product that they were using, so they were wanting to understand how they could use analytics to identify that.
So, these two use cases in mind, this center of excellence, they did have one key person in charge who was working with to evaluate the vendors and everything. They evaluated over 50 vendors and then that 50 went to 12 to 6 and then ultimately, to 1. They then deployed analytics to both of these use cases starting from reliability and then process. And it went really well. When they deployed the analytics solution onto the reliability, they were able to find a problem which their automation system was not able to because the problem was well within the alarm limit that it was intended to be. So, they were immediately able to change their maintenance practices to address that. They then moved on to their process use case and, of course, after that, now they are on the path to scale it up to multiple parts of the organization and multiple plants. So, that was a very good use case where they followed that whole step and now, they are in the process of expanding.
Jim: Yeah. That seems like the way to go about it, identify the possibilities, find the ones maybe with the best chance of success or greatest return opportunity, prove out the value with those and then look to scale and find more opportunity throughout that. Are there things people involved in these types of projects with analytics that they should watch out for and avoid or any kind of guidance you can give?
Manasi: One of the things that I have noticed that people do is they start first with technology and then they think about the use case. So, they would go about, just with an idea in mind that they want to deploy analytics, they would go about starting talking to vendors and looking at the details as to what that application does or does not without really taking a step back and understanding why they should be looking at it in the first place. So, that’s one thing that…setback that I have seen that has happened to them. So, something else that I noticed was getting involved…the timeline that they set for to deploy the use case is so far ahead that they get too involved in that use case itself rather than actually evaluating and giving short iterations and short successes. So, at that time, they end up deploying that use case for more than one year or more than two years without actually having to go back and see why they started this initiative in the first place. So, that’s one thing, starting with the technology itself and worrying about that rather than what they would want to apply to, I think.
Jim: Yeah. And that seems to happen, not specific to analytics but so many things. It’s almost like starting with the possible solution without really a full understanding of the problem, like I have the hammer, now let me go look for some nails or something, instead of understanding the problems, especially what’s the biggest one we can solve that’s not necessarily the hardest one to solve and then build from there.
Manasi: Yeah. That’s a good point. The other couple of points that just came to my mind was another one was working in silo. Let’s say the initiative is starting by one particular site and they’re not really in touch with their corporate, let’s say the IT guys or the technology guys to see what vision do they have. So, working in a silo and not collaboratively. The other key point for analytics to work, people are also important. They are the subject matter experts who are looking at these daily end-to-end results. So, it’s important to continue to invest on them and train them on the analytics application that they are working on. So, not investing there as well is where we have seen where they would decide to go with one application, let it run and then keep on running with that. It is important to have a roadmap. If you are deploying 12 models, out of which these 3 require tuning, retuning, going back and seeing whether it needs to be fitted well with the new sets of data or in a different way that the plant is operating. So, not having that vision in mind, just having one use case and then going with it, that was some of the pitfalls that we’ve seen.
Jim: Well, I think there’s definitely words of wisdom in there for people to look out for things not to do in this to make things be more successful. How can they learn about how our folks at Emerson involved in analytics can help?
Manasi: Emerson has a lot of information out there in our external website. One place that I keep pointing them onto is www.plantweb.com which has a lot of information on how to get in touch with some of our subject matter experts such as our Operational Certainty consultants, our solution architects, our industry experts. They certainly can help with some proven practices such as facilitated workshops or some benchmark processes. They can help you walk you through this process of understanding how to deploy or how to go about finding that use case or deploying analytics in your organization. So, get in touch with your local business representative who can help you get in touch with one of our consultants to help you walk you through that process.
Jim: Yeah, I think what you said there is really important to maybe help on that front end to really scope that process, maybe even bring that cross-functional team together on how to talk in a way that it’s productive and moving forward to identify what are the big opportunities there. Well, Manasi, thank you so much for joining us today. And you enlightened me and hopefully, for our listeners, you’ve enlightened them in to where to get started, things to consider, things maybe to avoid, and how to engage with Emerson to help them improve their business performance in all those different ways. So, thank you so much.
Manasi: Well, thank you Jim, and thank you for providing the opportunity to speak to you.
End of transcript
Visit the IIoT Analytics: Plantweb digital ecosystem section on Emerson.com for more on the array of operational analytics to help you improve overall business performance.