Predictive Maintenance – One Step at a Time
Predictive Maintenance: An Overview
Rather than waiting for failures to occur, predictive maintenance (PdM) allows you to identify likely failures before they occur, significantly reducing downtime. This sounds excellent in theory, but what does it mean for you? How will it benefit you? What will your organization need to do in order employ predictive maintenance and achieve the desired results?
Benefits of PdM
A recent study by the Aberdeen Group showed a drastic performance difference between Best-In-Class maintenance organizations (top 20%) and the laggards (lower 30%). Top performers boast the following performance scores:
- 1.7% Unscheduled Asset Downtime
- 91% Overall Equipment Effectiveness (OEE)
- 20% Return on Assets vs. Corporate Plan
- 31% Reduction in Maintenance Costs
On the other hand, lower performers report the following metrics:
- 14.8% Unscheduled Asset Downtime
- 73% Overall Equipment Effectiveness (OEE)
- -11% Return on Assets vs. Corporate Plan
- 0% Reduction in Maintenance Costs
Downtime, as you know, can be extremely costly. A 2006 auto industry study proposed that stopped production costs average $22,000 per minute – $1.3 million per hour! A separate Dunn & Bradstreet report calculated that, for a 10,000 employee company, labor costs alone total $896,000 per week of downtime. Your organization may be more resilient, but it cannot be argued that downtime is expensive. Consider the wide-reaching impact of a downtime event:
- Lost labor time
- OEM & other emergency contract labor
- Tooling costs
- Replacement parts and rush shipments
- Workaround (band-aid) implementation costs
- Start-up costs upon recovery
- Reduced customer/investor confidence
Predictive maintenance monitors equipment and environmental conditions and uses analytics to identify high-risk situations before they deteriorate to a downtime situation, significantly reducing maintenance and operational costs.
Getting Started with Predictive Maintenance
The above-referenced Aberdeen study identified several strategic actions that were implemented by Best-in-Class maintenance organizations. These include using analytics to plan capital expenses as well as outsourcing and scheduling non-critical maintenance before it is needed. Rather than introducing a sudden culture shock, such measures may be implemented in a step-by-step manner, allowing the organization to grow into a mature maintenance framework.
The Foundation: Visibility & Control
As an asset moves through each phase of its life-cycle, it is managed by different groups, such as project planners, engineering, maintenance, or operations. Appropriate workflow processes, whether manual or automated – or somewhere in-between, allow each group to obtain the appropriate level of control and report accurate asset details while it is within their control.
Likewise, each group needs visibility into the asset data that was collected during each stage of the process – from planning and procurement all the way to its current state – in order to make intelligent decisions.
As a first step, take a look at your asset management system and ensure that (1) the appropriate users have the ability to update asset data as they make changes, and (2) all affected decision-makers, regardless of which part of the cycle they own, have visibility into the entire process.
Establish Basic Metrics
Next, establish basic metrics to identify where your organization stands now, and how it improves over time with each successive process improvement. Many, many different Key Performance Indicators (KPIs) have been suggested for maintenance measurement. However, it is far more valuable to select a small number of KPIs that are important to your organization than to display a large list of numbers that may have limited value.
When implementing predictive maintenance, a few downtime-related KPIs may be particularly useful, including:
- Number of breakdowns
- Percentage of unscheduled downtime
- Overall Equipment Effectiveness (OEE)
- Mean Time Between Failures (MTBF)
- Mean Time to Repair (MTTR)
Each KPI has its own relative strengths and weaknesses, which will be discussed in a future article. However, by choosing just a few – or perhaps just one or two metrics that are particularly relevant to your strategic goals, you can identify changes that either help or hinder your maintenance goals.
Implement Condition Monitoring
Condition monitoring is often under-utilized, but it is a key component of predictive maintenance. Just as with KPIs, start simple. Select a small number of critical assets to monitor. Next, determine what tests may be conducted to determine the equipment’s current operating condition. OEM documentation is often a good starting point for such detail. For example, to test a transformer, leakage current, insulation resistance, and temperature may be monitored over time. When a parameter is outside of an acceptable range, action can then be taken to inspect and or correct problems on a scheduled basis instead of waiting for failure at an inopportune time.
Automate Data Collection
Once monitoring points and their corrective actions have been established, the next logical step is to automate data collection. Your organization likely obtains data from various meters, sensors, and SCADA systems. Your monitoring system(s) may even be integration-ready, requiring only a small amount of effort to share data with Maximo. Integrating these systems with Maximo allows important data to be shared between systems in real-time, eliminating delays due to user interaction and inaccuracy due to user error. Several tools are available to do this, and will be discussed in future articles.
Once the foundation of predictive maintenance has been established, you may benefit from the sophisticated analytics provided by IBM’s new Predictive Maintenance and Quality solution. It combines your asset data, sensor input, and weighted KPIs to identify and manage potential failures on a large scale.
Implementing a predictive maintenance program will help your organization achieve significant savings by reducing downtime and maintaining customer confidence. Such a program may be implemented in a step-by-step manner, allowing the organization to achieve continuous improvement in a controlled manner.