With companies increasingly looking to achieve high levels of reliability and uptime, they must transition into a much smarter method of maintenance. Overall Equipment Effectiveness (OEE) is a metric used in the manufacturing process, which considers the various components to measure overall effectiveness of equipment. On average, the OEE at which companies should be operating is around 98 per cent; however, some are operating as low as 80 per cent. Generally, the UK is on average operating at an OEE of below 90 per cent.
These figures highlight the urgent need for companies to address the way they carry out their maintenance. Millions of pounds’ worth of savings could be had by increasing an OEE by just a few per cent.
First of all, can you tell me about the strategy of smart maintenance?
Traditionally, you still have two approaches. The corrective maintenance when something is broken or in a fault condition: this is scheduled maintenance. Then there’s condition-based maintenance where you try to identify when a failure is about to occur and do something about it before it happens, and you get the benefit. I can save the cost of the scheduled maintenance, if that was going to be undertaken unnecessarily.
Smart maintenance is having that ability to do this condition-based maintenance in a more effective way, so you can start to predict what’s happening. You can work collaboratively to achieve this. If you don’t have the capability on your own site or within your own engineering team to do that condition assessment, you can search the internet, send them data, send that information to experts elsewhere. They can carry out that assessment and provide recommendations.
What is the process of implementing smart maintenance when you visit a client?
You do a practice assessment. On the one hand, you look at the data, you look at KPIs, such as the maintenance schedule appearance, and OEE. We look at the cost of maintenance, but then you’d also look at the practices. You look at the procedures that they had. How well they were documented. How well they were being implemented. Things like training and development. How well they use suppliers. Whether they do things like assessment of spares, for streamlining the number of spares they have got. You have the chaos phase and the control phase. Then you have got the systematised phase and the optimised phase.
I think a lot of companies are in the systematised phase and then they acquired a maintenance management system, but they’ve not embedded the full use of it.
Recently, I was looking through some maintenance records for an oil and gas company that has 12,000 notifications of maintenance in its SAP system that need doing. Then I got another list of hundreds of anomalies. These systems don’t talk to each other, so there is a manual exercise there, of going through and trying to find out whether all of the issues that have been identified have actually been actioned.
That is the next step in smart maintenance; it is not the Internet of Things, but just having the discipline to use consistent name conventions and filling all the boxes you get in all these databases, so that you can more effectively manage maintenance in the future.
Some organisations still seem to have a run-to-failure strategy. Is that a valid methodology, given the modern tools available?
You need to do a criticality assessment. You need to look at what the health and safety and the business consequences of failure are. You just get your asset inventory of all your items and you’d start that way, to say: ‘What are the ones that are really critical? We really need to put some force and effort into getting that right. Which are the minor items that, if those are going to fail, that’ll be fine?’
For critical items, you either decide to put some redundancy in or, if you are going to be reactive and you can’t afford to put in redundancy, then you need to make sure that the maintenance strategy there is going to minimise the downtime and maximise the ability to be able to operate between failures. This means making sure that, if you do need to lubricate, the specification of lubrication is as it should be.
Most unsuccessful maintenance strategies that I have witnessed fail because of lack of engagement. How can this be overcome?
You must try to instil a culture. My colleagues did a project over about two years of workshops. They started with a big bang event on reliability, and how and why reliability is important, and that involved lots of people, lots of disciplines. Then they followed this up with some workshops. After, they have had ongoing cosy chats where they simply listen to people’s concerns and ensure that it is something that people have time in the diary to talk about.
Culture is probably the final step on the permanent reliability path. Usually, it attracts technical and systemic issues. In the end, you probably go to people, try to make sure that everyone is involved, from maintenance to purchasing to HR, and that everyone is safe. It needs to be done in parallel.
You need to work on the culture at the top with the management, and you need to do the hands-on competence and procedures that hit the bottom level. Then, you need to target the enabling level for the systems at the middle. You need technician-level, engineer-level, and management-level input as well to make sure that they do not action implementing change at the same time. You need them all involved in any programme so that you do get the elements of the pyramid, the culture change at the top of the pyramid and then all the support and detail that goes underneath to support that.
With the growth of third-party supplier maintenance, how hard is it for onsite maintenance staff to remain engaged?
It makes it harder to keep control of competence. As an example, we are building a collaborative operations centre for a customer. The idea is that we rationalise.Rather than have an ABB technician on all their assets, we have a central control room where they can look after all of the control and safety systems centrally.
Now, the problem there is that the operators’ technicians will lose that day-to-day contact. Then, you maybe need to start formalising channels of communication and training sessions, so that they don’t lose sight of what is needed to ensure operation and maintenance or control and safety system operation maintenance.
The danger is the competence gets fractured and you lose the continuity between the technical knowledge and the asset knowledge. There might be some special considerations that go on for an asset. If you are a product specialist, you might be drafted in to a set site and not be aware of the special conditions for that site. I think that’s the danger that can arise.
You experience silos in a different way. So, you will have the technical expertise, but you will need to get that balance you’re looking for, but also the asset in the site knowledge. Whereas if you’ve got an asset team, you might not have the specialist product knowledge, but you have the site knowledge, so it’s finding that balance.
Are connected devices and IOT the solution to some of the maintenance problems?
The danger is that, if you look at adherence to maintenance routines, some companies struggle. They have real trouble just doing scheduled and corrective maintenance. When it comes to condition-based monitoring, the same problems arise. If it is a recommendation of end-of-life failure and it gets ignored one week. Then, if the machine doesn’t break, the following week the same notification will come up and the reaction might be, ‘it didn’t break last week, it is safe to ignore it’.
You really need to think about that criticality exercise. Which machines are the most critical ones, if they do fail. Perhaps the second phase is the vulnerability exercise – so which items are most likely to fail – and really focus there. If you decide sensors are cheap and the internet makes it possible to connect with IPv6, so that you connect billions of devices to the internet, the danger is you are going to create more information than the organisation can cope with. That is the risk that you create so much noise that you make the maintenance people fall back on each mission and they mess things up.
It was the same way back in the Sixties. The control and safety systems were all hardwired relay-based systems. Alarms were expensive and process design had really thought about alarms, when and which systems were critical in alarms, but then in the Eighties and Nineties, with SCADA systems and DCS systems, alarms became cheap and you could put alarms on everything. You ended up with the human factor issues, in that there were so many alarms happening that the operators were unable to work with them.
Then alarm management became a discipline, trying to manage the disability and the criticality of alarms, and making sure that there was a manageable level of control through technicians. These are the sort of things safety and maintenance technicians will need, to make sure that we don’t overwhelm them with information.
How are disruptive technologies such as machine learning and AI impacting maintenance strategy?
We have not got that advanced yet. We are looking to drive that technology, but we are not ready yet. If you can have an engineered model or physical systems model, then you can potentially have a better understanding.
I think it would be interesting to try machine learning technology, because it is expensive to build and maintain engineering models in a virtual systems model or simulating models. It’s much easier just to collect the data and process it.
The worry is when you are trying to implement that in the field: will it be trusted, will it be used, will it be improved as time goes on?
Given that, what are your recommendations when it comes to maintenance today?
It’s important to think about these tools to help operations and maintenance, and you need to keep them in control to some degree. If it is something that gets imposed from a board, then it is likely to get distrust, at worst; then it is ignored, rather than being something that people own and improve.
The key is to get people involved when you first need things done and make sure that, before you do try to implement a clever optimisation strategy, you need to; that you establish the business case to implement these systems. Have you taken the basic steps? Do you have your maintenance management systems properly documented? Do you have the computer systems in place that people are using?
Because, unless you’ve got those basics, you will not be able to implement the recommendations that these machine algorithms and smart maintenance techniques are trying to tell you.
BOX OUT: The importance of proactive maintenance
You wouldn’t dream of embarking on a long car journey without checking your tyres, oil and fuel, yet many plant managers fail to regularly check their equipment, which is why proactive, rather than reactive, maintenance is key to reducing downtime.
In the food industry, machines are often used around the clock, with little flexibility for unscheduled downtime. This means that the machines must operate at optimal performance and plant managers need to be well aware of the condition of the machines.
Cost of stoppages
“When moving items around the factory, whether through bucket elevators, conveyors or vibratory feeders, it is important that these machines are in working order or else production may stop entirely,” explains Stephen Harding, managing director of materials management specialist Gough Engineering. “Despite this, many plant managers fail to schedule in planned shutdowns to allow maintenance to take place.”
Often, maintenance is carried out after a machine has broken down. However, every minute of halted production is costly, particularly in the food industry. Having a reactive maintenance policy, rather than a proactive one, means that there is a forced shutdown and it is not good practice.
“Therefore, even when considering purchasing new materials handling equipment, plant managers must consider how they plan to maintain it,” Gough adds. “An important step in this process is choosing the right supplier to source the product from, where the experience of the manufacturer can be trusted.
“Having an inventory of operational spare parts can also make it easier for trained engineers in the plant to quickly repair the broken part. Without this safety stock, the entire plant could come to a stop for the sake of a single component.
“While engineers from Gough Engineering are available to come and repair parts, having scheduled shutdowns for maintenance means that they can be there as a planned production intervention to minimise disruption, rather than waiting for a specialist to travel through the night.“
In the food industry, it is inevitable that parts will start to wear out from long service. In this case, by factoring in planned shutdowns, regularly scheduled maintenance visits to the site give specialists the opportunity to spot and replace wearing parts before they suddenly break and stop production.
“Food plants also risk accidental damage when machines are dismantled and reassembled for cleaning to comply with hygiene regulations," Gough continues. "This is often done during the evening by occasional contract staff that haven’t been trained effectively in how to do it.
“Often, the machines may not be put back correctly, or components are dropped or damaged, leading to breakages or poorly functioning equipment.
After every period of cleaning, equipment should be checked prior to reinstating production; although this does add slightly to the shutdown time, it may avoid extended periods of unplanned shutdown.”
With materials management machines, such as conveyors and sieves, forming the backbone of the factory, production cannot continue if they should come to a halt, points out Gough Engineering’s Harding.
It is therefore vital that plant engineers consider the total cost of ownership of their equipment and invest in items that come from a reputable supplier, with an effective and frequent servicing programme.