What Is a Lifetime Variability Curve (LVC)?

Use data to drive smarter reliability decisions at your facility with Quantitative Reliability Optimization (QRO). QRO is a new dynamic reliability approach that bridges existing first principles reliability models with new data science, multi-variate analysis and system-based optimization to drive improved facility performance balancing availability, process safety, and spending performance.

QRO is comprised of four elements: Asset Risk Analysis (ARA), the Lifetime Variability Curve (LVC), system view, and reliability simulation and performance optimization. The second component of QRO is the LVC.

An LVC is a dynamic model that predicts asset failure through the application of data science principles. Leveraging statistical distribution, the LVC can quantify the uncertainty around a probability of failure (POF) curve helping facility leaders make more accurate decisions.

How Does an LVC Work?

LVCs work similar to hurricane trackers – as the model collects more data, the model updates and better predicts when assets will fail. LVCs can be used for both functional failures and asset degradation.

An LVC model begins with a known data point either modelled or measured. It then creates a statistical distribution from this known point and incorporates other data such as process data to generate a potential range of predicted failure points. This distribution is then related directly to the probability of failure curve for the asset. As new data enters the model, the model adjusts to provide a more accurate representation of probability of failure for any given asset.

Example of an LVC

For example, an LVC can be applied to the vibration of a bearing in a centrifugal pump. The LVC helps predict when the vibration of the bearing will cause the pump to stop functioning.

In Figure 1, the LVC leverages existing vibration measurements to predict a cone of uncertainty of when the bearing will fail. Since there is not a particular trend with existing data points noted, the model predicts a wide band of uncertainty.

As the LVC is fed more data, the model continues to refine itself, quantifying a POF curve. As a result, the model is able to create a more accurate band of uncertainty as seen in Figure 2.

Benefits of an LVC

With the incorporation of an LVC, facilities are able to:

  • Leverage a probability of failure curve that is continually refined based on facility data instead of a static probability of failure
  • The ability to collect the right data at the right time, helping facilities justify spending and better identify the optimal time to perform maintenance activities relative to planned downtimes
  • Incorporate process data into the probability of failure curve and quantify the impact of process excursions and their impact of asset life
  • Blend data science and subject matter expertise to help facilities avoid over-collecting data in some areas and under collect in other areas

To learn more about how an LVC fits into Quantitative Reliability Optimization (QRO), read about QRO here.

More resources like this