CMMS technology may be becoming increasingly sophisticated, but at the same time the whole maintenance process is benefiting from a smart, accessible and flexible approach, as Michelle Knott discovers.
We are at a turning point when it comes to putting predictive maintenance into practice. The idea of deploying smart systems to let engineers know when an asset is likely to fail is not in itself new. But what’s changed today is that innovative technologies and techniques are bringing what was once an expensive, labourintensive process within the reach of more and more process operators.
In addition, those companies that might previously have managed only their most critical assets in this way will find it rewarding to extend those capabilities to equipment throughout their production sites.
Asset performance management (APM) and computerised maintenance management software (CMMS) systems are wellestablished in some applications, but a range of recent and emerging developments such as increased connectivity and the Internet of Things, cloud computing, advanced data analytics, machine learning and artificial intelligence are together facilitating a step change in the way such systems can inform decision-making, both in terms of maintenance and more broadly.
So there may not be more data but [new capabilities] mean that it’s going to enable users to make better decisions than they have in the past
Dan O’Brien, business and strategy director, Honeywell Connected Plant
“Innovative technologies are changing the landscape of how our customers are able to measure and improve asset performance – no question,” says Dan O’Brien, business and strategy director for Honeywell Connected Plant.
He stresses that success is as much to do with how plant data can be used as it is about gathering extra data: “Customers have already spent a lot of money collecting instrument measurements and asset data. My contention is that 75% of that is not used to make data-driven, actionable decisions. So there may not be more data but [new capabilities] mean that it’s going to enable users to make better decisions than they have in the past.”
“There will definitely be much more data available and, as always, the smart thing is what you do with it. It’s about drilling in to get the kind of insight that will allow you to make good decisions,” confirms Jon Moody, chief product officer with SSG Insight, which offers APM based on its Agility software. After more than 30 years of operating as SoftSols Group, the company was rebranded as SSG Insight at the end of last year to reflect the way the industry is transforming.”
Moody adds: “We’re working with a number of clients who are already seeing big gains thanks to greater use of analytics. The real benefit in having the data is starting to build up that predictive model. It’s one thing to say that a piece of equipment is outside of a threshold, but you need then to use intelligence (and that’s increasingly artificial intelligence) to say when do we need to act on that? Based on historical evidence, what do we need to do and when? That all flows through into other things like stock management and enables you to manage the whole maintenance process in a much more efficient way.”
Ronald Thorburn [pictured] is responsible for global analyst relations with supply chain planning and optimisation software company Quintiq. He adds that extra data can feed into maintenance decisions that impact the business in a range of ways: “The value isn’t in the data. It is what you are able to do with the data. How you can make it actionable. If a vital piece of equipment does go down unexpectedly, Quintiq can provide real-time decision support in terms of calculating the effects of the delay for all the affected orders.
“When scheduling production, it can take into account all planned maintenance and is flexible enough if it can work within a time range for the maintenance, it can optimise when [assets] should be taken off line and still maximise production. With our growing use of machine learning and predictive analytics impacting forecasting, this technology can be combined with our optimisation techniques to create a maintenance plan that brings greater efficiency to the way manufacturing operations perform.”
A major driver of this emerging revolution is that systems are becoming more straightforward to implement and use at the same time as they are becoming more sophisticated. “Think about a computer in 1985 and the expertise you had to have to install software or send an email or send a fax or connect an attachment. Now every five year old or 80 year old on the planet is pretty good at using a smartphone and doing sophisticated computational exercises that people with a masters in computer science would have been challenged by 20 years ago,” says O’Brien.
“So what we want to do is connect to more assets and serve that up in an easy to understand set of dashboards or key performance metrics or even personalised notifications to someone’s smartphone. Honeywell is going to create an analytics-driven asset performance management environment that chases you down to give you information, versus you having to go hunting for it.”
“Over time, machines and their sensors will employ higher levels of artificial intelligence, such that they won’t just warn or predict, but will, in some cases, be able to even fix themselves,” agrees Thorburn. “Flexibility is also key for the solution to be able to both adapt to a specific industry and to grow with the needs of the organisation.”
Analytics is the loudest buzzword in this new environment, where it makes it much easier to extract nuggets of useful information from a mountain of data. However, there are other essential ingredients that must also be in place, such as data security. “Secure data connectivity is not new but it’s got a level of criticality today that no one could have imagined it would have ten years ago… We thought about security for a long time but now it’s front and centre in our consciousness,” says O’Brien.
The value isn’t in the data. It is what you are able to do with the data. How you can make it actionable.
Ronald Thorburn, global analyst relations, Quintiq
Conversely, openness is critical to enable APM systems to access useful data from external sources. Moody says that a truly open architecture across the industry would be the ideal, but that’s unlikely to happen. However, he thinks that the existing community of middleware developers will help smooth the way: “There is already a whole industry of middleware out there so it’s not such a barrier as it might have been in the past. There are standards that mean that one system can talk to another with only a moderate amount of development and there are also cloud-based interface layers.”
Openness should also help users to access data from beyond their own industrial systems. For example, Moody suggests that weather data can be incorporated into predictive models by looking at historical patterns: “The other thing we’re looking at is external factors, so there might be external inputs into the predictive model. So, for example, can you see trends in the kinds of breakdowns or issues that occur in terms of weather?”
Saving, not spending
SSG is scheduled to add a business intelligence tool (Agility BI) to its suite of solutions in May to make it easier to carry out this kind of trendspotting and ‘joined up thinking’.
So will this brave new world mean that current CMMS users will have to rip up existing systems and start again? Not according to O’Brien: “First, we want to understand the investments our customers have already made in collecting data. What are the wireless structures that they’ve put in place? What historian are they using and what data from their critical assets is already being collected? What condition monitoring systems do they have in place? Most customers that already have a plant up and running are doing something, so we always try to minimise the amount of capital investment a customer has to make.
“We can start working with the data they already have and we’re often able to bring our customers a lot of valuable insight without a single new piece of hardware, simply by collecting, moving, organising and applying some of our asset framework models and insight calculations or analytics to that existing data.”