Five guiding principles can help equipment OEMs and industrial equipment users to design a successful data-driven maintenance transformation that extracts full value from the data their equipment generates, argues Adrian Bostock
Today’s industrial environment is a warehouse of data, with many pieces of equipment cranking out vast amounts of information about the way they have been working thanks to a growing suite of sensors and infrastructure that captures and collects data.
Most companies do a good job of collecting this data to figure out what happened, but then what? Many original equipment manufacturers (OEMs) are incorporating data capture and collection capabilities in their equipment and others are complementing this architecture with data repositories and advanced analytics layers.
Despite the access to these data, to extract full value from them – in particular reducing Total Cost of Ownership (TCO) and the impact of unplanned downtime – big steps still need to be made. Similarly, among their industrial customers there is a wide range of end users.
Early adopters have identified which value drivers matter most and use sensors to define the maintenance regime that minimises TCO and losses, whereas other companies with large legacy asset pools have not started assessing the value that can be extracted from the data or investing in pilots to discover and scale the data-driven benefits. In many cases, maintenance is still having to jump in to put out fires when they pop up.
So, why focus on data describing the past when the value from data lies in the future operations? This is where machine learning (ML) and artificial intelligence (AI) come in: ML/AI can discover patterns leading to failures based on many historical data that the human cannot detect due to the large number of variables that affect the health of the equipment.
Leading OEMs and their industrial customers use these techniques to predict failures and prescribe solutions, for example to optimise replacement budget allocation or to avoid expensive repair jobs. By combining the predicted likelihood of failure, cost of failure and the costs of preventative work, a balance between reactive and proactive maintenance can be struck to minimise the sum of TCO and costs of failure.
Despite these analytical advancements, most industrial companies are still struggling to get to grips with TCO and costs of failure which are driven by unplanned downtime, poor throughput, late deliveries, excessive inventory, and quality issues that elevate operating costs and compromise sales.
Their reactive approach towards maintenance – which in contrast to an analytics based approach does not provide clarity on the root causes of failures and downtime – has led to lack of accountability & authority over the performance of their industrial equipment while the complexity costs of multiple maintenance contractors has been increasing.
It’s time to take back control over the equipment
For end users to be competitive in their industries, they need to be able to accurately predict how their equipment will perform in the future — not just how it has performed in the past. State-of-the-art predictive maintenance can address this challenge and create three powerful benefits.
First is a lower total cost of ownership (TCO) without compromising asset availability and throughput. ML and AI can process the wealth of data that comes from today’s industrial equipment — creating information that can be used to predict and maintain the asset’s health and allowing inspections, maintenance, replacements and repairs to be planned at a time that minimises the impact on the equipment’s availability or throughput.
Secondly – a greater stability for operations and supply. Every company wants their equipment to be as healthy as possible. When this is the case, production is more predictable, more customer orders can be delivered on time, fewer spare parts are needed, and there are fewer rush orders. And finally, better relationships with suppliers, customers, and employees.
The benefits of more stable operations move beyond the four walls of the factory with greater customer satisfaction and better relationships with equipment and maintenance service suppliers, something which is increasingly important given today’s growing scarcity of skilled engineers. In addition, more stable assets means that employees can spend more time finding ways to improve the equipment’s performance rather than looking back to figure out why it failed — creating a culture of innovation.
Five ways to launch a data-driven maintenance strategy
Making the switch from traditional hypothesis-based reactive maintenance to data-driven predictive maintenance is no simple task. Most organisations launch a pilot programme with a handful of assets, but few manage to scale it to capture the full value across all sites.
- Focus on creating value and identifying the most important assets. Start with a value driver tree for operations, and determine which assets are most important when it comes to factors such as total cost of ownership and avoiding unplanned downtime.
- Assess the granularity and robustness of your data. Is the data from your equipment actually giving you information that you can use to make different decisions? For example, does it differentiate between planned maintenance activities and the reactive ones? Is the data collected in a way that is unbiased?
- Iterate, refine, and scale. Think big, start small, and scale fast. Take an agile approach, using a multidisciplinary team that can develop resilient solutions. Take small steps with many iterations. Striving for perfection from day one will almost certainly end in failure.
- Engage stakeholders at all levels. The most successful digital and analytics transformations are a partnership that is led from the top—the very top— and executed by the shop floor.
- Improve your data architecture. Have models, policies, rules, and standards that govern which data is collected, how it is arranged and integrated, and where it is used.
Applying these principles will increase the chances of success for your data-driven journey towards a competitive position with respect to TCO, supply chain stability and relationships. It is a journey, not a project, which will embed a new way of working that will require a different mindset, skillset and toolset. It will allow your organisation to focus on improving the future performance, a characteristic that will attract the best talent.
Article was co-authored by Adrian Bostock, Principal, London ([email protected]), Antti Kautovaara, Partner, London (antti[email protected]) and Ernst van Duijn, Partner, Amsterdam ([email protected])
For more info visit Kearney
For more from The Manufacturer click here
All images courtesy of Kearney