Erik: Welcome to the Industrial IoT Spotlight, your number one spot for insight from industrial IoT thought leaders who are transforming businesses today with your host, Erik Walenza.
This podcast is brought to you by PTC and Live Works, the world's most respected digital transformation conference. In this series, we will feature partners of PTC who are driving digital innovation and value creation today. I'm your host, Erik Walenza, CEO of IoT ONE. And I'm joined today by Preston Johnson. Preston is the platform lead for intelligent monitoring and industrial IoT for the company Allied Reliability. And Allied Reliability is a subject matter expert that provides consulting around predictive maintenance.
So this conversation for me was very interesting today, because Preston is really at the top of the game in terms of expertise around the use case of predictive maintenance. He has more than 30 years of experience in this space. We dove deep into the thought process behind people processes and technology control, predictive maintenance solutions, where we're going, what challenges you might face, what results you might be able to achieve. I hope you enjoy our conversation. And thank you, PTC, for helping to pull this conversation together. We appreciate your support. Preston, good morning.
Preston: Good morning. It's a pleasure to be here.
Erik: Yeah. Well, thank you for joining. I do appreciate the time. Preston, where you calling in from today?
Preston: I'm, actually, in New Orleans, Louisiana.
Erik: What are you doing in New Orleans or New Orleans?
Preston: We're going to see if we can help the water and sanitation district come from a run to failure environment to a reliable low cost environment.
Erik: And I imagine in New Orleans, I would imagine it's a challenging one just because they do have at least a history of storms. Is that impacted? Or is it more or less like any other American city?
Preston: I think it's a typical most American cities. We've helped some other water sanitation districts and other cities go from where New Orleans seems to be from phone conversations to a reliable operation. So I wouldn't say it's uncharacteristic.
Erik: And where are you there? Are you just kicking off now?
Preston: We'll be doing a lunch and learn to introduce the team to Reliability best practices: people, process and technology in technology, ultimately, will be contained an introduction to the industrial Internet of Things. We've got to help them get their people in process, legs of the three legged stool together as well.
Erik: So, Preston, we're going to dive into probably a deeper discussion here of predictive maintenance. But why don't we kick off just with a little bit of background on yourself and on Allied Reliability to give a foundation for the conversation? So I know that you've been with Allied for a while, but what where's your background coming from?
Preston: Yeah, so I'm an engineer by education, an information systems guy by a graduate school, 28 years with an industrial instrumentation manufacturer, went from 100 employees to 7,500 employees at that particular company, who's still a partner of Allied Reliability’s. I actually recruited Allied Reliability as a partner for my former employer, and then just so walked across the aisle to continue my own journey and understanding the use of technology for business benefit in the manufacturing space.
So I've been with Allied Reliability for three and a quarter years, primarily played a role in helping them adopt the latest and greatest technology for condition monitoring, and predictive maintenance. So it's been a lot of fun helping to coach and mentor the traditional teams in the use of connected sensors, connected technologies, and the industrial Internet of Things.
Erik: I see on LinkedIn, your title now is “Platform lead, intelligent monitoring and IIoT”, I've got to imagine that that evolved over time. Or three and a half years ago, was Allied Reliability already thinking in terms of IIoT and was this position existing at that time?
Preston: The position did not really exist at that time. What we were trying to do is simply go from traditional human-based data doing automated data gathering perspective, that has historically been my claim to fame and bringing new technologies to industry. So my role really was in a sales role to help our clients understand there were less expensive ways to collect condition monitoring data, data that will indicate the reliability and the capability of the machines.
And ultimately, we evolved into the industrial Internet of Things as the little S curve of technology adoption started to take off. And several players came into the market offering platforms that allow us to leverage our domain expertise from our past 25 years of experience in the industry, and leverage other technologies and hardware and software technologies that exist, and a single stack, if you will. So my role really is to help organize and help clients understand what those components are, as they think about an industrial Internet of Things or a digital strategy for their particular manufacturing or plant operations.
Erik: There's one other interesting point on your CV, which is that you're the Vice Chair for these Society for Machinery Failure Prevention Technology. I can tell that you, let's say, for lack of a better way, geek out on this. So I can see that this is a really a passion of yours. How did you end up running as the Vice Chair for this society?
Preston: Well, my former employer, we were bringing some technologies from the test and measurement world into the industrial space. And I managed a group of product managers and R&D teams to take the technologies that were being used in test and measurement and use them in the industrial space. I needed a forum to share how those would work.
So I started building and attending and participating in smaller conferences that were centered around condition monitoring. And the Society for Machine Failure Prevention Technology was a nice, small community with some excellent domain expertise from a number of members. It's been around for 50 odd years, started out as a power and maintenance interest group that was really spun out of the Department of Defense, joined up with a vibration Institute. So currently today, the Society for Machine Failure Prevention Technology is a division of the vibration Institute. And we're beginning to work a little more closely with them to integrate our organizations. We're actually having our annual conference the week of May 13 in Philadelphia,
Erik: The annual conference, is it open? If I happen to be in town at some point, would it be possible for somebody to attend one of the meetings? Or is this more of a closed discussion?
Preston: No, it's an open conference. And yes, you'd be welcome.
Erik: Before we really dive into the use case, let's cover a little bit more Allied Reliability’s perspective on this and where they fit into the game. So can you just in, 30 seconds, 60 seconds give us a quick understanding of how Allied Reliability fits into the topic of predictive maintenance?
Preston: Well, we look at predictive maintenance as a mechanism for helping our clients get the most productivity out of the equipment. So in a nutshell, we work with asset intensive industries who are interested in maximum productivity of the equipment that they employ and utilize. We've been doing this for 25 years.
Our founder is a founder of a Society of Maintenance and Reliability Professionals, which really have built out a series of body of knowledge that drives best practice behaviors in operating, manufacturing and process industry equipment. And those best practice behaviors drive best practice results. So in Allied Reliability, we have a reliability consulting team and a condition monitoring predictive maintenance team.
So there are elements in the reliability organization such as planning and scheduling, inspection of materials, inspection of equipment, lubrication management, and all of those work together to drive the productivity at manage repairs and maintenance activities on the equipment that we've seen reduced annual maintenance spend upwards of 20 and 30%, in some cases, reduce spare parts inventory by 30%. In some cases, and in one particular case that had multiple plants and made a beverage product, they were going to have to build another plant to meet their demand. And we found that hidden plant amongst three or four of their existing plants by just increasing the productivity of the equipment and aligning their personnel to implement best practice behaviors.
So we take on reliability consulting as a first leg, help them get their organization in place, and then we help them improve their equipment maintenance plans, so they're more reliable and operate at less maintenance costs.
Erik: So Preston, predictive maintenance is one of the use cases that is certainly most commonly referenced today, or it's one of the first use cases that companies explore in the context of IIoT because of the fairly clear path to ROI. But how do you view predictive maintenance in the larger context of digital transformation that companies are undergoing?
Preston: And digital transformation certainly does cover a lot of aspects that an organization might encounter. And when you look at some of the market studies, LNS Research for example, they list quite a few of the different digital transformation efforts, everything from customer optimization, getting them to spend more on the particular application, or products into process optimization, which might come in as a second to predictive maintenance.
One of the interesting elements is process optimization is certainly a, an excellent tool for understanding that pressures and flows and chemical reactions and so forth in the process, optimizing work in progress and buffers between machinery work cells. But if the equipment is not reliable, if it fails unexpectedly, if its performance is chaotic, then it's hard to optimize a process with those chaotic variables within them.
So, predictive maintenance breeds the stability of the manufacturing or the chemical processes that then allow that digital transformation and process optimization to start to take place. So I think that's why predictive maintenance and equipment reliability, and even the energy that the equipment consumes tend to show up at the top of digital transformation return on investment.
Erik: When you were typically working on a project, is one of these usually the driver for value creation? Or is it a combination? Because I suppose, when a company's deciding how to invest in their building out their business plan or behind the investment, some of these factors are probably fairly easy to project the impact, somewhat more complex to project. And my understanding is that you would be part of the discussion around a work could be the potential impact. So how would you typically walk a customer through prioritizing areas of impact in building out the business case for an investment in a predictive maintenance solution?
Preston: So there's a number of ways of measuring the uptime and the value of the product that's produced by the machine when it's running. And then downtime is loss of revenue. And then another aspect of maintenance cost and cost of maintaining and operating the equipment, we oftentimes like to look at the cost of production per unit produced. And if we can lower the cost of production per unit produced, machinery reliability has a significant way in both at uptime. More uptime, the more products we produce, the more revenue we make, the lower the cost of maintenance, and operation of the machine, the lower the cost per unit, and ultimately, those go together to increase profitability for the facility or the plant.
There are other aspects that might be a little more subjective, or a little more soft: safety, unreliable machinery can be unsafe; environmental, if we're mixing the wrong, we're not cleaning the water before we return it to the system correctly, that could have environmental impact and a large industrial discharge wastewater could apply environmental impacts, goodwill, fines, and so forth.
And then lastly, product quality. The best operation of the equipment leads to the opportunity to operate the equipment with the best raw materials in the best way and produce the highest quality product. So if we're producing scrap product, or product that we have to sell at a lower cost at a lower feature set, there's some opportunity in the quality perspective as well. And when we help our clients understand those returns on investments, we have to find existing data in their existing accounting systems or maintenance cost systems so that we can show that we've got a trend of historical data, and that we're going to be able to move that in the right direction often the case of revenue down in the case of costs, for example.
Erik: In your experience, is that always possible? Do you, as part of your business model, have a success factor where you need to be able to benchmark and show a significant improvement? Or is it often more of a project base where you do a particular piece of work, and you track results, but you might not be able to benchmark and show that there is actually a significant improvement because the benchmark did not exist?
Preston: Certainly. So, since we've done reliability work and condition monitoring work over the past 25 years with over 1,400 companies across 16 industry verticals, we do have some pretty good benchmarks where we've had success in driving the uptime, and lowering the maintenance costs of the equipment with those clients that we work with. So we have some internal benchmarks, society, maintenance and reliability, certainly provides some best in class benchmarks on some of the various measures that we can use to measure reliability, and financial metrics.
So there are benchmarks that we can utilize to measure. Where the challenge often comes into with a digital transformation, we often times start small and try to scale fast. And in that starting small phase, we may have a third element, which is to fully understand the technology gap. We do want to make an impact and move the needle somewhat on some of these financial metrics to show that we're actually starting to make an impact. But those financial metrics may be small enough that they really don't make a true return on investment, as you would expect from a larger project.
But as we provide the reliability engineering consulting, we usually try to help the clients understand where their numbers are, trying to find the low hanging fruit, what can we help them improve that is going to give them a return on investment so we can take that return in investment invested back into the same reliability journey that we're helping the clients on so that we can take the next level of fruit to fruition?
Erik: But based on your experience, this 1,400 success cases, what might be a range? I mean, you mentioned this one situation where as a beverage manufacturer was able to basically create or identify the productivity of a factory among three or four existing factories. So that's we're looking at a 25% improvement in productivity there, that seems like at the upper end. But what would be a typical expectation for improvement in uptime or reduce maintenance costs in a situation or maybe a range that seems feasible?
Preston: So, for some organizations, if we're helping them drive improvement in uptime and productivity and as low as 2-5% improvements, the volume of product that they're producing produces a large financial return. In one particular case, we're helping an organization understand energy consumption. Their machinery may still have some issues. We're taking a look at utilizing condition monitoring, and predictive maintenance as part of a digital strategy in a food and beverage manufacturer.
But they also have recognized that energy consumption is something they've really never measured. So we found some technologies that are going to help us measure energy consumption down to individual lines. Walking through their facility, we found conveyors that just run for an hour or two when there's no production upstream. We found other machines and lights on. And so there's a number of just getting data visible to the organization can help them make obvious decisions that can help them lower their cost of operation, lower the cost of their product and be more competitive.
So, energy might be one particular aspect, just helping them be more efficient in their maintenance processes, more lower their overtime costs for crafts skills, increase their wrench time, the amount of time that they're actually spending doing maintenance work as compared to finding all the parts and instructions and so forth, and perhaps even improving their production rate or production count. There's a number of embedded metrics that make up the uptime and the cost of production to leverage the revenue and the cost of production revenue measures, spare parts inventory, backlog of work orders.
So we can work on a number of these smaller bleeding indicators that can help drive the overall larger numbers at the end of the financial period, and help the organizations improve those. That really just depends on where that low hanging fruit might be, and how far we are from the end goal of improving that aspect of a maintenance and reliability program.
Erik: Let's talk a bit about the technology. So in your experience, where is the state of the art in predictive maintenance today? And what are the technologies that are driving progress?
Preston: Well, one of the first elements of the Industrial Internet of Things is just the ease of connectivity of the software vendors that make industrial machinery connectivity tools, bringing the data to the business side of the organization, not just showing it up on the human machine interface at the machine. OB Foundation coming out with the universal architecture for connectivity, MQTTS is another communications protocol that really allows these Industrial Internet of Things platforms to pull all of the existing data from the control systems, while maintaining the security of the control systems and the reliability to control systems just to make that data much more visible.
And then when we need to add new sensors to get another process parameter or condition indicating parameter, a health indicator of the equipment, there's a number of wireless technologies that have recently come to bear, Bluetooth, different versions of WiFi, that are making it easier to add additional sensors without having to run additional cabling or make room for them within the existing control system. So those are two elements of connectivity that have really started to enable the Industrial Internet of Things and the state of the art.
On the software side, there's been a number of big growth in companies and tools or big data analytics, taking massive amounts of data, organizing those into events, or clusters or patterns, natural language processing, and we're already starting to see the advent of augmented reality. You can think of heads up safety glasses, heads up dashboards, and some automobiles that are kind of luxury automobiles today. But those same heads of displays are becoming available, and heads up safety glasses in the near future. And so I think we're going to see a lot of the information flow back down to the operations and maintenance teams through augmented reality.
So there's a lot of software and the connected worker that are bringing things together. So it's connectivity and big data software and connectivity with the enterprise that I think are the state of the art technologies that are really making the digital transformation possible.
Erik: I know you're not a system integrator, you are a subject matter expert. How do you interact with these technologies? How closely do you have to understand and help your customer make decisions regarding when new sensors need to be installed or decisions between software selection?
Preston: So it's really boils back down to the historical laws of physics and laws of the business process. So in the case of laws of physics, we look at all of the equipment from a failure modes analysis perspective. What mechanisms might the equipment failed to perform its function? Is there a sensor that can detect a defect as soon as it arrives and whose defects will cause that machine or equipment to fail to perform its function as sometime in the future? If we can apply that sensor, we can see the defect early in its life, and allow the organization to plan and schedule removal of its maintenance activity sometime in the future when it's operationally convenient to perform the maintenance. And that way the organization can plan and schedule the maintenance activities at the organization's convenience versus the machine deciding when it's just going to stop working.
On the software side, there's a lot of business processes understanding the maintenance and reliability process. And Allied Reliability works with a number of organizations to help them improve their computerized maintenance management systems, which typically orchestrate maintenance works, have all the job plans for common maintenance activities, plans and schedules those jobs, works with a craft skill schedules and so forth. And integrating those computerized maintenance management systems into an overall digital strategy brings historical work order close outs, costs, causes, and so forth together as well.
So the software on the computerized maintenance management systems, the historians that help us bring together, process data as well as condition monitoring data so we can see the patterns there and then all the new sensor data really helps us bring together just a better dashboard of the interplay of the people in process on a computerized maintenance management system side, the operational context from all the process data, and then the condition monitoring data that can help us detect the defects that caused the failures. So it's our domain expertise that helps orchestrate that.
Erik: And we discussed in our pre-discussion that one of the strengths of Allied Reliability is your database of knowledge gained over past projects. And so you have basically this understanding of machine dynamics and the dynamics behind maintenance event. This, you mentioned was embedded in the ThingWorx platform. Is that platform focused for you or the use case specifically on maintaining this database of knowledge gained over past projects? Or is it also used to integrate that together with data over, let's say, live projects or real time data coming from client’s systems?
Preston: As a condition monitoring services organization, when we deploy our over 100, 200 condition monitoring subject matter experts, we want them all to perform the condition monitoring services, data collection, data interpretation, and repair recommendations that they see from the data and a common high quality mechanism. So we developed a series of standards that allow us to collect the data correctly, to interpret the data correctly, and to make the right repair maintenance recommendations on that particular data. So that's one big element of our databases are operational standards for collecting, interpreting and recommending.
The second element is when we help our clients set up an equipment maintenance plan, sort of upstream a little bit in the reliability consulting arena, we look at the cause of failures of equipment. And while we've seen over 4 million pieces of machinery, and today, we analyze upwards of 30,000-50,000 components a day. So we're looking at motors and pumps and gearboxes all the time.
We collected the likely and probable failure modes for all of those components, which ultimately make up those 4,000 machines. And it's our asset health matrix library that as we call it that really lists out the common components across 4 million name machines that we've seen, how those components what parts they're made up of how those parts can fail, what causes those parts to fail, and what sensors can detect the defects in those parts. And that's ultimately our world class maintenance library component that we've built into the ThingWorx platform, allowing our clients to easily see oh, I've got a motor, a pump, a gearbox, what are the likely probable failure modes? Do I have sensors that can detect defects that caused those failures? And if not, if I added them, how would that help me be protected from surprised failure?
Erik: To what degree are you able today to project the impact if we installed X number of sensors in this domain, this is the impact? To what extent is simulation feasible using the existing data versus experimentation through a pilot project? Are we still primarily reliant on doing pilots and using common sense? Or are you able to, in many cases, really simulate the potential impact of an investment of sensors or some other technology?
Preston: From our domain expertise, when we're adding the appropriate sensors to components and machines that we've seen in our past in our database, we have a high competence that if we see a defect, we'll see it with vibration or motor currents or oil analysis will be able to catch that defect upwards of three months before it causes the failure of the machine and be able to make that repair recommendation.
When we make a repair recommendation, we describe the severity of the defect that we've discovered. We describe it as low, high or critical. In other words, low, you have a few months; high, you have a few weeks; critical, you have a few days. So it's the severity, how much the data screams out the presence of the defect that we're able to have categorized in our databases, in our operational standards to allow us to share with the client how to mitigate this particular defect. So all of that is based on domain expertise, and the laws of physics that make the machines operate.
Erik: When you're working with a new client, how do you determine whether the client is ready for predictive maintenance deployment? So if we think about the greater context of technology that's installed in a factory, let's use, maybe a factory in a third tier city in China where I'm currently sitting, for example, many factories have very little sensors, very limited capability of the people in terms of IT or integration, then so it doesn't mean necessarily that you can't using Bluetooth or WiFi, you can't install some wireless sensors, and maybe begin to isolate specific machine and at least identify maybe a solution to a specific piece of bottleneck, technology or equipment. But it does mean that type of factory has a lot of work to do before they're going to deploy maybe a more comprehensive predictive maintenance solution. So what would your analysis be when you're working with a new client to understand are they ready, or what type of solution are they ready for given their status quo?
Preston: How do our clients initiate with us the opportunity to A, improve their uptime, lower their maintenance costs and progress towards a digital strategy? And how do we help them understand where they are, and help them to develop a roadmap, or what we call a journey towards a more reliable plan to get them to some future state?
Well, the first thing that we have to understand is there what I'll call a three legged stool, there's people, you have the right people and craft skills and the right positions: there's processes. Do you have a mechanism to identify maintenance tasks, work management system that helps you prioritize and manage those work tasks? And do you have technology that can help you gather equipment health and equipment productivity data?
So we try to start with helping the client to assess where they are, and the people in the process and technology help them build a roadmap that leads them into a digital strategy that ultimately allows them to perform their predictive maintenance, improve their reliability, at lower and lower costs, by improving the reliability, gathering those reliability gains financially, but also lowering the costs of running that reliability program with the use of technology. So in some cases, where clients really are running their equipment to failure, they really don't have a good program in place, we really tried to stand up the people in the process elements.
We've been into one account where we stood up the technology leg too soon, and we were reporting defects and hundreds of equipment and explaining what the maintenance activity was supposed to be. But the planning and scheduling and the craft skills and the work management just weren't in place. We actually got yelled at for telling them at all if their equipment had defects. So we have to be careful and balanced and make sure that the people and processes are in place to be able to manage the data and do something with the data that we're gathering from the equipment, whether it be productivity data of the equipment or condition indicating data from the equipment.
So we tend to start with an assessment, help them build that plan forward using people process and technology legs of the stool, try to build those out equally on the journey to helping that client be best in class and reliability.
Erik: But often, companies tend to focus on the technology perhaps because it's somewhat simpler to think through purchasing a new technology to fixing a problem as opposed to trying to change people issues or process issues. And other, obviously, if you tell people that we're going to change or new skill sets are required or we're changing processes, then there's some sense of threat in cases. Situations where a company wants to focus in deploying new technology but you strongly feel that before that happens you need to address people in process issues, how do you have that conversation?
Preston: Well, I'd like to start out with letting them kind of like, an efficient story, you hook a fish, and you sort of let them run for a little bit, and help them make some progress in that direction. We can start with a technology implementation and just be cautious that maybe we're going to have to come back and work on some people and process.
In a particular case, in food and beverage, they feel like that they're running their plant very reliable, they feel like they're making their snacks at a profitable level. They want to use digital transformation to help them optimize the process, and understand the health of their equipment, although they claim they have very high uptime, as compared to their scheduled downtime.
So, they added hundreds of sensors and technology, we turned on our ThingWorx application, and about 40% of their equipment showed signs of defects. And now we're kind of helping them recognize that yeah, those defects may already be there. And cases where we've gone into, started a predictive maintenance program with other companies that have never really tried to detect defects and very run to failure-oriented, typically, we will see defects and a large significant of their equipment. It just hasn't failed yet. It hasn't started screaming. The operators haven't noticed hot to the touch and loud noises from the machine.
But the sensors are seeing the defects much earlier on the point of detection, the point of failure curve and that's what technology does for us. So we have to kind of back up and help them understand how the condition monitoring technology works, and how we should best leverage that information and perhaps prioritize that equipment based on criticality and importance to production.
So it's a balancing act. Certainly, we want the client to be able to lead and take ownership of the overall reliability process. Because at the end of the day, when we get them down their reliability journey, we're helping them hire and take ownership of the technologies and the practices. We continue to offer services and a lot of companies take advantage of our services. But a lot of companies want to internalize those capabilities as well. So we have to encourage them to internalize where they are interested in it, and help them learn by doing.
Erik: So there's going to be maybe a pilot or an early stage, and then you'll scale from there. For initial value delivery and then more of a full scale deployment, what would be looking at in terms of timelines?
Preston: We can start to see some results, some of the needles that we're measuring move in as early as a few months, 2-6 months. The larger paybacks when we start to use the financial net present value, internal rate of return calculations on what we've done, typically start to show themselves in about 18 months, sometimes as early as within a year. Sometimes it might take a couple of years, just depending on how visible the numbers are, and whether we've picked the right low hanging fruit.
Erik: We've already covered the some of the benefits, then more quantitative and the qualitative benefits. What about the risks? So I suppose there's predictive maintenance technology, it doesn't necessarily have an inherent risk in itself. But there's certainly risks that the project fails or that the project doesn't arrive at the expected benefit. What would you identify as the areas that a company should really pay attention to as they start planning to make sure that they don't neglect an important issue that comes back to bite them as the project matures?
Preston: So there could be a number of risks if there's a significant investment in technology. Without the people in the process, you end up coming out with a loss for the project. You put a lot of money in but you weren't able to utilize the data to implement the uptime or to lower cost of production activities, so risk became a loss of money. You put in a bunch of cool stuff, maybe a few folks enjoyed working with that cool stuff. But in the end, they didn't produce any monetary value. So there's a financial risk there as well.
There can be other risks. For example, where we're going to throw in all this technology, therefore, we don't need as many people out walking around looking at the machines. We are going to just repurpose those to some other role within the organization, and automatically all of this new data that we're going to collect, this data tell us how to optimize our process right away, it's going to tell us how to maintain our equipment right away.
And if we don't include domain experts that understand the laws of physics and the chemical processes, the mechanical processes, the electrical processes, there's a risk of over dependency on the technology without understanding the physical laws that are in place within the facility. And we could have risks associated with unexpected failures of equipment. The machine big data analysis just hasn't learned from the data just yet, or is giving observations that don't drive action. And we're expecting that we put in this security blanket, but the security blanket just isn't well grounded in the principles of physics. So there's a risk of over-reliance on the technology without a full understanding of the subject matter: the laws of physics of the process: chemical, mechanical and electrical.
Erik: Are there any concerns here about cybersecurity breaches as you start to potentially to connect a closed facility to external data sources?
Preston: There is, and that's always a really interesting challenge. And that's why the Industrial Internet of Things and digital transformation is really a team sport, requires leadership from the organization, from up in the higher levels of the organization to help get the various departments aligned, both the operations, maintenance reliability, as well as the information technology team.
And the information technology teams, if they're aligned, and they understand the journey and the benefits that the organization is trying to achieve with their journey, then they will be very useful in isolating networks, making sure that data flowing is leaving the facility to go to a cloud or outside the facility that it's not data that could be used against them, that the firewalls are in place yet the data can still flow through them. The information technology team can be a real asset to a digital transformation, as long as they're excited and motivated to participate in. And they're the ones that ultimately are going to bring up the security issues. But if they're motivated to be a team player, they'll help the organization navigate those security issues in a way that security is covered, as well as functionality is covered.
Erik: Preston, I do want to be cognizant of your time. But I would love to dive into one or two specific case studies, if you have time for that. Would there be one or two case studies that you have in mind that you'd be able to walk us through?
Preston: Sure. So this is a case study that's public, and it's been in the power generation and electrical power industry news for a couple of years now. And that's the largest power generation company in the United States, that's Duke Energy. They have had some failures of equipment that cost them to fail to meet their production requirements or commitments. It raised their costs of electricity. They had to go buy from their competitors, plus they had some major failures and major costs of equipment.
And as they wanted to improve their predictive maintenance program, they took it upon themselves to recognize that we need our subject matter experts; instead of walking around collecting data, we need them analyzing data and telling us something that's going to happen in the future so that we can better plan.
So, Duke Energy sort of had their reliability team internally, that was before I joined Allied Reliability, but they took it on themselves to replace manpower, walking around collecting data, observing machines, with technologies. And my former employer actually was able to provide them with enough sensors and data acquisition systems to build their data flow stack. They built the information technologists were involved. They developed what they called a maintenance network for all of the vibration and temperature and oil sensing data, motor current sensing data, thermography data to flow to a historian. They utilize the OSI soft high historian.
And then they added some maintenance and reliability tools from the Electrical Power Research Institute plant view, and some machine learning tools from a company, InStep, who has a product by the name of Prism. And they merged a lot of their expertise into their corporate headquarters in Charlotte, North Carolina. And that's where their maintenance and reliability room is with a mixture of data scientists, subject matter experts, and those that are advising the individual plants who are taking advantage of this data. So that's been a real success.
They've documented millions of dollars of annual savings, and they're actually making better use of their outage planning because they have condition indicating data that tells them what work needs to be done on their machines at each of their next outages. So, that was my first hurrah in putting all that together. I was very much associated with the sensor and the data acquisition technology and the communication of the information to the subject matter experts and to the historian in that particular case,
Taking those same types of knowledge and experience, and allied reliability also gets involved in heavy metals, heavy industries. In one particular case, we get involved with steel and aluminum manufacturers and those organizations operate in a different types of environments, typically around the Rust Belt, where they have different classes of labor that we might find different than in the southeast where Duke Energy is. In other words, we have a union labor, we have craft skills, we have contractors. So the organizational aspects are a little bit interesting.
But the heavy metals industries in the Americas is trying to be more and more competitive. Certainly, there's tariffs now that are helping a bit. But they still need to improve the reliability of their equipment. In one organization, we spent a year and a half helping them understand what assets they have in their facilities. So we did some reliability consulting right up front. And then we're able to help a small group which is the rolling press and the hydraulic pressure pumping systems. They were just upgrading those, so we took that opportunity to deploy a digital transformation element around that small scope of one particular group of machines.
They took the ThingWorx platform and are actually bringing in process data to let their managers of each of the individual processes see all of the product counts and product flow. But down at that particular machine level, when the machine was actually commissioned, the Industrial Internet of Things, the condition monitoring sensors, and the process sensors that were all flowing into the historian started to show some anomalies that the new equipment was starting to have. And they were able to quickly react to those anomalies before they became bigger problems, and further optimize the hydraulic systems that are driving the role stands.
The operators of the roll stands are experiencing much smoother operation, and the roll stands because of the smoother more in control, more reliable hydraulic system. So that's a success in potentially reliability and product throughput and in identifying defects before they became a bigger problem.
Erik: And both of these sound like pretty straightforward solutions. Any challenges that you encountered across either of these projects, or were they both pretty smooth deployments?
Preston: Well, there's always challenges. When you mix a number of software technologies on a network, there's integration between those two. Sometimes the integration effort comes from an internal team. And sometimes the integration effort may come from external skilled resources. So we try to work with our clients to help them understand that we're going to need some integration between the data acquisition systems, the historian and the computerized maintenance management systems. We're going to need some dashboards for the leadership who are interested in the metrics and the KPIs and the needles and the gauges that we're trying to move.
And all of these are really kind of out of scope of allied reliabilities competence, capabilities and core value. As we're really reliability consultants, we understand the people, the process and what causes machines fail. We help our clients implement equipment maintenance plans, what sensors how to interpret the sensor data, how to make repair recommendations. But the rest of the organization has other systems, and we're not systems integrators. We’re domain expertise, and we know a lot about how the technology works and how it integrates. But we expect others to do that.
So if we don't have a systems integration team that's lockstep with us, we'll have little bumps and bruises along our path. Little speed bumps where the systems aren't talking to each other, data is not flowing, it's getting dropped or what have you. And as long as we're keeping that three legged stool balanced, we're in good shape. If we don't balance the three legged stool, then we may be a little tilted when we sit at the table.
Erik: Well, maybe that's a good ending point for our discussion: take care of your people, your processes and your technology. Preston, any last points that you'd like to make?
Preston: Yes. So I'm wholeheartedly believer that digital transformation and the industrial asset intensive industries around the world are something that's very popular, very interested amongst many organizations and very possible with the technologies. We just have to think of it as taking step by step. Don't bite off more than we can chew at any particular time. Think about our three legged stool: people, process and technology. And think about building a strong team that starts with the organization's leadership, brings in domain experts that understand the processes and the machines, the laws of physics. And then add in the information technologies, augmented reality, machine learning, networking, communication, cybersecurity, and make sure that we've got a good team of systems integration partners that can help us put all the pieces together and deliver the business and productivity results that we're expecting.
Erik: Well, Preston, only one more question from my side. What is the best way for listeners to get in touch with either you or your team or to learn more about Allied Reliability?
Preston: The best way is our website, alliedreliability.com. And there there's opportunities to drive into various aspects, so, people, process, and technology, request information, and our sales and consultants will be easily available. So that's alliedreliability.com.
Erik: We'll put that in the show notes as well. So Preston, really appreciate your time today. You have an absolute wealth of knowledge. Thank you for taking the time to share with us.
Preston: It's been very much my pleasure, Erik.
Erik: This episode of the industrial IoT spotlight podcast is part of a collaboration with PTC, the global software company that helps companies design, manufacture, operate, and service things for a smart connected world. To learn more about PTC, visit www.ptc.com and to collaborate on future podcasts with IoTONE, please feel free to reach out to us at team@IoTone.com
Thanks for tuning in to another edition of the industrial IoT spotlight. Don't forget to follow us on Twitter at IotoneHQ, and to check out our database of case studies on IoTONE.com. If you have unique insight or a project deployment story to share, we'd love to feature you on a future edition. Write us at erik.walenza@IoTone.com.