The path to prescriptive maintenance
The Industrial Internet of Things (IIoT) will bring many benefits as it evolves and advances in the coming years. But today many companies worry about their readiness and feel overwhelmed thinking about the expense of preparations, particularly those organizations that are still moving from reactive and preventive to predictive maintenance (PdM). Now the topic is prescriptive maintenance, where analytics can show that a piece of equipment is headed for trouble and can prescribe prioritized, pre-determined, expert-driven mitigation or repair.
These ideas can be overwhelming, but in this case, as in most cases, planning rather than worry saves the day. Rather than attempting a heroic leap from reactive to prescriptive maintenance, reliability organizations and companies can perform many activities today to prepare for the IIoT and for prescriptive maintenance while simultaneously finding immediate reductions in maintenance costs and increased availability.
One of the most commonly cited benefits of the IIoT is the elimination of downtime, a direct improvement on production and profitability. But the benefit only comes as companies move away from old habits such as reactive or preventive maintenance and are prepared to act on detailed information that predicts asset failure.
Prepare to be proactive
Imagine receiving notice that an asset will fail in the next 10 days and creating a corrective work order to address it only to have the work order linger in the system until the asset fails! It happens to many companies that use PdM. Prescriptive maintenance may only increase the precision and frequency of the information in this scenario, therefore, until the organization is prepared to act on alert information in a timely manner, little benefit will be realized.
Three actions will move your organization to PdM and will lay the groundwork for prescriptive maintenance, and for your adoption of IIoT initiatives when the time is right.
- Prepare your culture to be proactive
- Integrate your condition monitoring program with your maintenance work management processes
- Implement a continuous improvement process.
Top performing facilities enjoy consistently strong reliability. In fact, top performers have high mechanical availability and low maintenance costs simultaneously (Fig. 1). Lower performing manufacturers seem always to be fighting the most attention-grabbing issue, while process availability suffers.
The business advantages of top-quartile performance—higher availability that results in lower maintenance costs—should be constantly on the lips of executives. Management support is the most important ingredient in transforming culture from reactive to proactive, and a well-documented business case is its foundation. To obtain approval, make it clear that these reliability and maintenance benefits come from PdM and that the gains profit your organization.
The secret to the achieving the goals of Fig. 1 lies in the illustration in Fig. 2. Because most failures occur at random equipment age intervals, a time-based preventive maintenance activity can permit an unplanned shutdown, resulting in more downtime than if the plant were using a predictive approach. Hard failures of assets are often more expensive to repair due to fees associated with obtaining parts on short notice, collateral damage to the asset, and the costs of scheduling resources on short notice.
Many facilities practice time-based maintenance in which at some pre-set interval a technician works on an asset to avoid a failure. As illustrated above, many time-based tasks are ineffective, and, in many cases, periodic parts replacement actually induces failure. Because the preventive approach requires more planned downtime and results in more unplanned downtime, facilities that follow this model face extreme challenges in attaining top-quartile performance.
Nevertheless, the preventive maintenance model has been widely employed since the 1950s. Changing the mindsets of operators, maintenance personnel, and business leaders must be part of every plan to attain PdM and top-quartile performance. The key idea is to instill in your organization the fundamental belief that failure is unacceptable and that everyone shares in eliminating its causes. Once this idea begins to take hold and stability (instead of firefighting) is rewarded, you can unlock the benefits of condition monitoring and adopt upgraded behaviors.
Take a holistic vision
To prepare your processes to attain positive results, first think about your asset management efforts in a holistic fashion. The Reliability Value Chain (Fig. 3) helps you envision how the pieces fit together in a continuous flow. All disciplines are coordinated by a core of reliability strategies that optimize what data is collected about assets and how it is used to drive maintenance efficiency. The chain links elements in four categories: obtaining accurate data, translating data into information, gaining knowledge from information and determining actions based on knowledge.
Consider how you obtain data about your assets from maintenance procedures, process parameters and condition sensors. Then consider how you use that data around the circle to analyze asset health and eventually plan and schedule maintenance work.
The value chain illustrates how sensors mounted on equipment are used to monitor the ways in which an asset can fail—failure modes. From a reliability perspective, failure does not necessarily mean that the asset dramatically stops working in some way. An asset has failed once it stops contributing what is required of it for a process to operate as designed. Condition monitoring and predictive technology together can detect when failures start to occur, giving the organization time to plan an appropriate response.
Top-quartile performers in asset reliability have been using sensing technology, condition monitoring, in their operating models for decades. Condition monitoring is increasingly a method of choice because the cost of deploying sensors and condition indicators has fallen and analysis techniques are growing more sophisticated. With a well-engineered condition monitoring program, you can apply the right types and mix of sensors to observe your equipment in nearly real-time. Monitoring will provide the data necessary to understand the condition of your equipment and employ PdM. From that information, you can plan to thwart problems according to your facility’s priorities.
A process and culture change
The question, "What happens next?" will help you understand when you have integrated your predictive information to work management—even if you may have to ask the question two, three, or even four times. For example, if you notice that a vibration reading is high. What happens next? The reliability engineer checks differential pressure. What happens next? The reliability engineer delivers a report to the planner. What happens next? And so on.
When the final answer is, "The repair is executed and the asset is returned to service," you will know that integration is complete. While workflows that can be documented electronically have some big advantages, they are only needed as backup to an organization that is well-versed and comfortable operating within a defined process.
Culture and process change is so important that it is worth stressing. Plant culture must change to adapt to new data in new forms. The business will quickly lose patience with investing in predictive maintenance if new sensors are added and performance results do not change. Only when the organization can readily use asset health information should it seek new forms of information. Having clear line of sight along the sequence of "What happens next?" will help ensure that you are ready to realize the promised value of prescriptive maintenance offerings.
Set the stage for continuous improvement
Think of attaining PdM and top-quartile performance as a voyage and continuous improvement techniques as your daily mode of navigation. It is very unlikely that you will chart a straight course on this voyage. You will need to adapt as you go, as unforeseen and unanticipated conditions reveal themselves.
Your navigational framework consists of a solid fact base about your assets. Notably, you need to document every maintainable asset in your plant. You need a basic understanding of how your assets fail and of the risks posed by their failure. You do not need to know every detail to get started, but you need everyone to share in adding to your knowledge as you go. Once your failure risks are identified and ranked, your course should be set to eliminate or mitigate the greatest remaining risks first.
When you evaluate risk, be sure to consider the potential quality, environmental, safety, production, and maintenance cost consequences of failure. Generally, these categories encompass all the ways in which failure can affect the business, but they should be adapted to your circumstances. Engage operations, engineering, Health Safety and Environment, maintenance, supply chain, and finance in the development of your rankings. Their participation will help them understand and agree on priorities.
Document your strategies
Once you have identified and ranked your failure risks, documenting what you plan to do about your reliability strategies is essential to keeping the program on course. Initially you will encounter the unexpected, and having documented what you expected will help you communicate what needs to change and why.
While the process of documenting reliability strategies can be resource intensive, it pays for itself by ensuring that your scarce resources are systematically directed at failures with risks that are in line with the cost of the strategy.
There is a strong business case for the use of predictive technology. There is also a strong technical case. On average, sensors can detect 70% of failures across maintainable assets (Fig. 4). It is not a coincidence that top-quartile performers base a very high percentage of their reliability strategies on detection technologies. As detection capabilities continue to grow, the relevance of time-based tasks to maintenance will continue to fall.
By documenting your current strategies as you move toward PdM, you are setting the foundation for future prescriptive maintenance programs. You can develop and document the risk of failure from each asset, so your facility has a plan in place ready to mitigate failures. And you develop and document a menu of strategies that play out when an issue is predicted and a prescription is provided by the equipment, by analysis software, or by a consultant.
Navigating to top-quartile performance is a series of adjustments to your continuous improvement process. You will encounter unanticipated and unmitigated failures and strategies that did not perform as expected. Initially your reliability performance will vary as you encounter the largest problems in the design of your program. It is a critical period to anticipate and address through frequent, periodic meetings to review and improve program performance. Ideally, you will have the chance to include all the key stakeholders in the meetings: operations, engineering, supply chain and plant leadership.
Every success—such as eliminating bad actors—should be celebrated with the same group. Just as you would start every meeting with a report on number of days since a recordable safety event or loss time injury, reporting on the trend in unplanned downtime hours at every meeting is a great way to highlight improvement and drive awareness.
As you deploy more predictive technologies on more assets, you will discover latent failures to add to your maintenance backlog. These will drive a temporary need for additional resources to address them. As downtime from disruptive events and latent failures diminishes, your program will stabilize into a trend of continuously rising availability, driven by the elimination of progressively smaller defects and the release of maintenance resources from reactive work.
The desired mix of work will let you know when to expect smooth sailing (Fig. 5). As the percentage of resources applied to understanding asset health through both PdM and time-based PMs trend toward 15% of total hours, and the results of these activities approach 35% and 15% respectively, your asset availability and maintenance costs will become increasingly stable and predictable.
A rising percentage of requested work consisting of process changes and design modifications will indicate that the organization is proactively eliminating recurring causes of failure. Regardless of where you are in the journey to top-quartile performance, shrinking the share of reactive work is a strong sign of progress.
Smoothing the way forward
You don’t have to be in the top quartile of performers to get ready for prescriptive maintenance. Taking steps now to bring asset health information into maintenance work will bring benefits today, accelerate your realization of benefits from IIoT and prescriptive maintenance in the future, and ensure sustainability of your program for the long term. Staying the course in moving from preventive to predictive maintenance prepares you for prescriptive maintenance and other evolutions in reliability that can drive significant performance improvements.
Both goals are attained by continuous improvement. A strategy fostering continuous improvement will prepare your organization for advances in technology, for staying competitive, and for increasing profitability.
Will Goetz is vice president of business development and marketing for Emerson.
See additional stories from the Plant Engineering May 2017 cover story below.