Use Predictive To Get to Prescriptive
EP Editorial Staff | March 15, 2020
Preventive asset-management is a stepping stone to prescriptive practices.
By Paul Gogarty, ABB Asset Lifecycle Services
As we live and work in the era of the Industrial Internet of Things (IIoT), we get used to seeing constant innovation in technology and equipment, and the new capabilities this offers across industrial and process automation. With this comes new ideas and applications that will become part of our toolbox and vocabulary in the years ahead, if they haven’t already. Our approach to asset maintenance is a good example of this evolutionary arc.
Historically, preventive maintenance has been the prevailing practice across industry. With this approach, maintenance is conducted at set time intervals, or after a certain amount of usage, usually pre-determined by timelines set by the manufacturer. It is hoped that performing full or partial maintenance on an agreed timetable will prevent unscheduled downtime caused by machine breakdown or failure. This entails strict record keeping and careful planning and scheduling of maintenance actions. A key element of preventive maintenance is that it is conducted irrespective of whether the machinery is operating optimally and, therefore, whether maintenance is actually required.
Internet of Things technology is permitting manufacturers to use data, gathered through automated systems, to actually predict a failure and prescribe preventive actions based on historical data that takes into account the specifics of the individual process, product being made, and the environment in which it is being manufactured. This approach prevents imminent failures, greatly assists with scheduling, optimizes uptime, and reduces overall maintenance costs by applying maintenance resources at optimum times. The result is a positive impact on an enterprise’s profitability. This technology is a stepping stone to successively smarter and more automated maintenance practices that make effective use of human resources.
The predictive and prescriptive terms are often used interchangeably, as if they refer to the same thing. This is incorrect. While predictive maintenance is being adopted with increasing frequency, prescriptive is still very much in its early stages and a state to aspire to in the future.
Creating the conditions
Predictive maintenance is built on a solid base of condition monitoring. Condition monitoring uses sensors and field devices to gather information on machinery and provide the intelligence to gauge when the equipment actually needs maintenance, rather than relying on pre-arranged intervals. It operates in real time, generating up-to-date information on equipment status and operations, informed by regular or continuous assessment and anomaly detection.
The sensors and devices employed can be connected to the plant network and shared across an enterprise through IIoT technology. The data generated by sensors and devices can be hosted in an edge or cloud environment, which allows them to be combined with data coming from other production sites, or with enterprise-wide historical data, and analyzed for the type of maintenance that should be performed. Edge and cloud applications have been important drivers of predictive maintenance, as they create greater opportunities to harness, process, and use data obtained from a wide range of sources.
Across industrial automation, sensors and detectors have been in use for more than two decades. It was only when the computing horsepower became available to accelerate data analysis for practical and effective use, and the IT industry started to understand how to use the required analytics, that the potential emerged for employing it in industrial automation. Producers were able to use increasingly higher volumes of data on asset conditions to create deeper insight and better analytics for asset management.
In addition to the more efficient and precise maintenance schedule it allows, the operational benefits that predictive maintenance delivers to plant operations are a key part of its appeal. Timely maintenance helps industrial producers avoid machine failure and the ensuing costly downtime. Performing maintenance only when necessary also reduces costs, compared with compulsory maintenance at fixed intervals.
What’s more, compulsory maintenance can actually induce unnecessary failures due to human error, such as forgetting a part when reassembling the equipment, or failing to properly tighten fasteners. In these instances, it would have been better not to touch the equipment at all.
Advanced predictive maintenance can generate analytics that demonstrate that the equipment must be maintained to avoid failure, and also to provide additional information on the probability of failure and on predicted time to failure. This information enables even more accurate and effective maintenance scheduling across the asset base.
This specific knowledge empowers producers to schedule shutdowns with confidence, allowing them to extend times between maintenance where possible to maximize uptime.
It also enables them to plan exactly which maintenance activities to prioritize, and which to postpone, without risk of failure. They can group together maintenance actions and be proactive in addressing imminent issues, depending on the probability of failure in the foreseeable operational future. This means duration of shutdowns can also be reduced, providing more calendar time for production.
In addition to the advantages predictive maintenance delivers to plant operations and productivity, broader macro-level factors drive its adoption. Market pressures contribute to the demand for constantly higher production availability which, in turn, requires maximum asset availability. Timely and targeted maintenance helps deliver this.
Similarly, pressure to increase margins forces producers to wring out every last drop of production capacity while reducing maintenance and operational expenditures. Predictive maintenance is an effective way to achieve this.
Aging assets and slimmer capital budgets are pushing producers to find new and more-effective ways to extend asset life. Efficient asset management and maintenance ensures the business is maximizing the capital outlay made on their machinery.
How is prescriptive maintenance different and what new solutions will it offer?
Our understanding of our assets and their modes of operation is constantly expanding. Algorithms now recognize when machines enter failure modes and can identify the pre-conditions around them. These processes are automated, but they will often require a human being to make decisions and sanction any associated maintenance actions.
The aim is to use machine learning and artificial intelligence applications to understand and spot patterns that lead to failure modes. Then, moving beyond simple monitoring and recommending, these applications can be equipped to make their own decisions as to the next maintenance steps and to act on them automatically, inasmuch as this is possible in a physical universe. With enough intelligence in the system, the idea is that the plant can operate with less human interaction in terms of overview and correction when the plant’s assets encounter disruption or enter unexpected states.
This presents additional opportunities for agility and operation responsiveness that can have a positive impact on productivity and profitability. Autonomous control can be used to change how a plant is operating, based on equipment status and/or product demand and price in the marketplace. Sensors and devices will continuously collect data, compare it with other machines in the factory and broader data from the cloud, and then generate insights and actionable responses in real time.
Prescriptive maintenance will allow producers to operate a plant in its most efficient and profitable mode while responding to factors such as market demand and maximizing equipment-maintenance intervals.
In the future, embracing the possibilities of prescriptive maintenance will entail changes in infrastructure, management, and operations. Adopting predictive maintenance, which itself requires new systems and working methods to be utilized effectively, is a step in the right direction. A prescriptive approach will require further change if it is to be successfully employed and its benefits fully appreciated. Existing systems will need to be updated and disrupted for the prescriptive maintenance approach to work effectively. EP
Paul Gogarty is Product Manager for ABB Asset Lifecycle Services, Peterborough, UK. He is responsible for the development and harmonization of ABB installed-base management services, Enterprise Asset Management tools, and lifecycle cost estimation for ABB Industrial Automation.