Prescriptive Maintenance – the “Holy Grail” Needs IoT to Get Started
We all need to be able to “do more with less” – people, parts and budgets. For those of us in the public/non-profit sectors, this is how we have a chance of hitting our budgets and tackling deferred or backlogged maintenance. For the manufacturing and other for-profit sectors, this might mean the difference between achieving profitability. “Lean” and efficient maintenance operations is a helpful contributor to these initiatives.
The spectrum of maintenance activities spans from the neglectful “run-to-fail” all the way to the holy-grail of “prescriptive.”
- Run-to-Fail (R2F): Only for short-life, disposable, durable, non-maintainable, low-cap assets, etc.
- Preventive Maintenance (PM): “Every X” – calendar (every week/month/day, etc.) or usage (runtime hours, cycle count, etc.)
- Condition Based Maintenance (CBM): Continuously monitor assets to spot impending failure (pressure, temp, vibration, oil analysis, etc.)
- Predictive Maintenance (PdM): Monitoring assets during normal operation to proactively avert downtime. Aided by IoT (the ability for your assets and facilities to use sensors to tell you when they need maintenance), machine learning and artificial intelligence.
- Prescriptive Maintenance (RxM): PdM, but with suggestions on actions to take to fix/improve – more discussed below.
Pushing Up the Probability-Failure (P-F) Curve
How do all these progressive, improved maintenance strategies help us? Metaphorically, it pushes us up the Probability-Failure curve (P-F curve). The P-F curve is an excellent description to illustrate the power of the maintenance spectrum and how it improves asset performance, reliability and longevity.
Starting with the “design” (D), this is the point where the asset is brand-new from the vendor. The “installation” (I) phase is where the asset is calibrated, fine-tuned and optimized for production in your environment – probably the best that asset will ever operate. At this point the asset starts actual production – real-life use. It is very difficult to see or hear, but over time that asset will degrade – eventually hitting the “potential failure” (P). This is where that asset is starting to need maintenance.
Eventually, if that asset is not maintained properly, it will eventually hit the “failure” (F) point – this is where the asset is no longer able to perform as designed/intended. If continued neglect occurs, that asset can easily hit “catastrophic failure” – the worst of all situations.
The role of maintenance is to push up the P-F curve. The earlier you catch problems, the less likely for all the bad stuff: unhappy customers (staff, students, citizens or fellow workers), poor manufacturing production, wasted energy, and the profit-killer of unplanned downtime. The more advanced forms of maintenance – RxM and PdM will catch problems the earliest. Standard PMs might catch issues early, but less intelligently. R2F is the worst, defeats the whole purpose (except for R2F designed assets).
RxM is the Holy Grail of Maintenance
Although much of RxM is still off in the future, the promise of prescriptive maintenance allows you to not only catch maintenance issues early but utilizing technology that can theoretically catch them before they ever happen, suggesting (“prescribing”) a solution based on lots of data, historical analysis, machine learning and artificial intelligence. Sound complicated? Like all maintenance software, the tools to deliver RxM need to be easy to utilize by the humans performing the maintenance. Breaking this down into smaller parts shows that RxM requires several easy-to-start with components, that in combination will eventually deliver the value of RxM. You need to start your journey with the basics, including Computerized Maintenance Management System (CMMS) and the Internet of Things (IoT).
Let’s go through examples to highlight this future holy-grail: a properly functioning pumping facility within your organization is central to your operational success. This generic example uses pumps/systems.
The ineffective, run-to-fail example:
- Although newer installed pumps and related assets run fine, they don’t see much maintenance (preventive or corrective) at first, the team is too busy fire-fighting with older pumping systems.
- Occasional preventive maintenance is scheduled, but not always completed – often skipped as the team is too busy with other higher-priority work orders.
- Eventually these pumps need attention and hopefully that work is completed before operation-killing downtime or worse: budget-killing premature replacement or overhaul of neglected assets.
- Unfortunately, to many organizations still use this dated, profit/budget-killing lack-of-strategy. You are on the wrong end of that P-F curve.
The better, preventive-oriented condition-based-monitoring example:
- Same pump systems, but now the organization is serious about maintenance and incorporating a combination of preventive and condition-based-predictive maintenance strategies.
- The periodic calendar PM work orders allow techs to spot check, inspect, perform basic PMs and other tasks to prevent major issues from happening.
- Adding IoT sensors – for example, a wireless sensor that can easily be affixed to a pump to detect vibration and other metrics to determine maintenance-related anomalies (cavitation for example) is occurring.
- This is possible today! Tools such as Brightly’s Asset Essentials CMMS and Smart Assets IoT platform will help turn a reactive maintenance situation into a much more predictable – pushing up that P-F curve.
- The future – prescriptive maintenance utilizing lots of appropriate data, sensors, and AI:
- Same pump system – a modern CMMS and IoT sensors are in place driving lean and efficient operations.
- All appropriate data related to the pumps are recorded into history. This includes preventive maintenance, corrective maintenance, usage-related data and similar.
- Ongoing condition monitoring via IoT sensors diligently watching for any anomalies that indicate maintenance is needed. Since “out of tolerance” can be a relative phrase, the AI adjusts acceptable ranges for vibration given current operating conditions and usage.
- Forecasting demand of the pumping systems is essential to determining how hard the asset can be pushed. This is fine-tuned by taking in the historical maintenance records. A dynamic, intelligent monitoring of the asset can appropriately and accurately determine if operations are at risk and maintenance is needed.
- If maintenance is warranted, analyzing historical work on that pump (or similar pumps), vendor info and other intelligence will help “prescribe” the best actions needed.
There is still more work for the software/technology vendors to achieve true RxM, but it is NOT too early for operations professionals to start this journey.
Doing “more with less” (creating lean operations) is greatly aided with technology. RxM is the holy-grail, but most take baby-steps to achieve. These baby steps include finding and properly implementing a quality CMMS and starting to harness the power of IoT. This will give you a solid foundation in preventive and even some aspects of predictive maintenance, pushing up the P-F curve and enjoying the operational and financial benefits. A solid CMMS vendor will continue to harness the intelligence surrounding our operations, allowing for better insights and intelligent decision making. In time, these tools will become more intelligent getting us close to true RxM – the holy grail of maintenance!