What is condition monitoring?
Unplanned downtime is consistently identified as one of the top manufacturing issues. Condition monitoring can offer a fairly simple way to start addressing this issue and helps users become more proactive in addressing and preventing impending failures of critical equipment by using data to anticipate problems.
There are four common maintenance approaches: reactive, preventative, condition-based, and predictive. Each has different cost-benefit tradeoffs, and it may be appropriate to use multiple approaches depending on the range of equipment in a facility. In general, the reactive and preventative approaches have significant drawbacks when used on critical equipment and when unplanned downtime is a major concern.
Condition-based monitoring and predictive maintenance (which uses condition-based sensors, tools, and data) offer approaches that can proactively warn of impending failure and are especially relevant to important equipment.
- Reactive: “Run until it breaks” might be used on non-critical, low-value assets, but is highly risky to apply to important components, where costs of repair and costs of downtime are high.
- Preventative: “Maintain at regular intervals, whether the asset needs it or not” might be appropriate when failures are age-related, but it may be that costly maintenance is being done on assets that do not need it.
- Condition-based: “Monitor limits on relevant indicators” can address failures regardless of whether they are age-based or random and monitors changes in one or more potential failure indicators, such as vibration, temperature, current/voltage, pressure, etc.
- Predictive maintenance and analysis: Attempt to learn from machine performance (condition-based data) to predict failure.
Condition monitoring provides warnings about faults and makes it possible to schedule repairs without unplanned downtime and lost production. It focuses on using sensors to monitor the status and health of machines, plants, or individual components (bearings, motors, fans, etc.) and then transmitting this data to control and/or supervisory systems for analysis and action. Continuous condition monitoring aims to detect changes and anomalies and can help customers record long-term trends and statistical evaluation – an entry point into predictive maintenance and predictive analytics.
How condition monitoring works
Typically, as a failure progresses, different indicators emerge (vibration, temperature, change in pressure & flow, etc.), and monitoring these can allow a more proactive approach than reactive or predictive maintenance. The Potential-Functional (“P-F”) Curve provides an example of the lifecycle of a failure:
Warning and alarm limits for the selected indicator(s) are set and when the limits are reached action can be taken. The limits can be set based on recommendations from the equipment manufacturer, ISO 10816-3 guidelines, or test data gathered from the machine. Over time, the data gathered can be analyzed to modify the limits and can be used as the basis for predictive maintenance and analysis.
When an alarm is triggered the maintenance staff can investigate and address the issue in a proactive manner – whether a simple task such as lubrication or minor adjustment, or a more critical part replacement – generally with time to schedule the activity during a planned downtime, rather than in the middle of production.
How to get started
We suggest you start with a small pilot system:
- Perhaps use a demo system, portable, or temporary tool.
- Set the initial alarm/warning limits based on ISO standards, manufacturer recommendations, or experience with similar machines.
- Gather data and look for insights.
- Modify limits based on data and consider using predictive maintenance software/tools for deeper analysis.
- Create buy-in with maintenance teams and the leadership team.
- Document the positive impacts of the changes and discuss them often.
- Grow the system over time.
Once you are ready to expand, an article in Control Engineering magazine provides advice on a process we endorse, including:
- Conduct a criticality analysis: Which assets are most critical and have the most impact if they fail?
- Identify probable failures the asset will experience: How has it failed in the past? What has happened to similar equipment? Does the manufacturer have recommendations?
- Decide on the technology best suited to detect each failure mode: Do you need to monitor a device, machine, or complete facility? What are the most appropriate indicators and the sensors to detect them? Do you need continuous or one-time monitoring? Where is the data needed and what is the best way to transfer it?
- Trend and analyze the data to plan and execute maintenance actions at the most advantageous times: How will you visualize the data? Do you want to use software to do analysis for you? Are there obvious trends and conclusions to be made?
Getting started with condition monitoring can seem challenging and complicated. By starting small you can learn what does and doesn’t work and take a more proactive approach to maintenance as you spread condition monitoring throughout your facility.