Improved analytics functionality in WellMaster RMS

Recent releases boost analytics capabilities with various survival probability and MTTF calculation methods. Available now is, among other things, Weibull with burn-in filter, assumed wear out, and mixed distribution.

WellMaster RMS has seen some powerful analytics capabilities added over the last couple of months. By request from our clients, we have added a number of new calculations and visualizations that provide data trending and prediction. This allows the operator to save time by getting the results they are looking for directly from the application.

Better prediction with Weibull

One key feature added is the Weibull distribution curve fit, used for prediction based on the observed failure history. From before, WellMaster RMS showed the Kaplan Meier graph for the observed survival probability, with the optional exponential distribution as the predicting capability. As failure rates are rarely constant, it often does not fit very well with the actual data. When introducing the Weibull distribution with two parameters, we see a distribution that is closer to the observed data points, and allows for an, in most cases, more accurate prediction. This is particularly useful in RAM analysis where higher early life failure rates have an impact on maintenance and replacements plans.

 

Survival probability (Kaplan Meier), compared with 2-parameter Weibull and Exponential distribution
Survival probability (Kaplan Meier), compared with 2-parameter Weibull and Exponential distribution

To further improve the prediction accuracy, we have included a variety of configurable parameters as well. This includes a ‘Burn-in filter’ to remove infant mortality failures, and ‘Assumed wear out’ to define the expected wear out for the component. The ‘Assumed wear our’ makes a better prediction of MTTF when wear-out has not been registered in the observed dataset. This prevents the highly unrealistic MTTF numbers often seen in reliability data dossiers.

An alternative to the ‘Burn-in filter’ is applying the ‘Mixed Weibull’, which combines two Weibull curves into one, thus providing improved fit with Survival probability graph in both early and late life phases.

 

Survival probability and Mixed Weibull
Survival probability and Mixed Weibull

In addition to the visual representations, WellMaster RMS provides the user with all parameters derived in the calculations. Depending on estimation method selected (Maximum Likelihood Estimation (MLE), or Least Squares Estimation (LSE)), the system provides the MTTF, as well as Weibull scale and shape parameters with confidence limits.

Check out the new features by logging in to https://wellmaster.exprosoft.com. If you do not have access yet, contact us at sales@exprosoft.com.