- Implemented a backend architecture to allow for customizations to be loaded and managed per tenant and facility.
- Removed the “Is Added” and “Is Removed” flags from the Asset Risk Analysis (ARA) screen.
- Allowed for multiple CML Readings to be recorded per day.
Probability of Failure Models
We have fully implemented the Probability of Failure (PoF) Model feature. Now you are able to drive your calculations in Newton by choosing the PoF Model you want to run, and the data you want to run with it. Here is a list of models available to use in this release.
- Creep - Uses historical pressure and temperature data, thickness measurements, and material properties to calculate projected remaining life as a function of time.
- Thinning (Linear Degradation) - Uses historical thickness measurements and modeled corrosion rates to develop a linear trend of expected degradation with quantified uncertainty in thickness measurements as well as uncertainty in actual degradation.
- Cracking (API 581) - Uses cracking susceptibility, inspection effectiveness, and other inspection data to calculate Management Systems Factor Score, Generic Failure Frequency, and finally Probability of Failure based on API 581 calculations.
- Vibration (Damage Onset) - Uses historical vibration data and Weibull to calculate the probability of failure.
- Oil Degradation (Damage Onset) - Utilizes historical oil analysis data and Weibull parameters to model the probability of failure.
- Weibull - Uses industry default or user inputted Eta, Beta, and Gamma values to predict probability of failure.
ARA Results Screen Changes
We have begun making updates to the results screen with more changes to follow. Here's an overview of what we have done so far.
- Simplified the KPIs that are shown at the Asset level
- Grouped the Asset and Component results
- Removed failure date lines from the LVC for CMLs that never fail
- Now showing planned downtime on the LVC results
- Fixed an issue where the Facility Administrator was not able to edit resource data.
- Fixed an issue where the Simulated Lowest Thickness Point was not normalizing readings correctly for cladded assets.
- Updated the Gibbs Sampler in the “Thinning (Linear Degradation) PoF Model”. This will result in the thinning models not having overly tight fits given the uncertainty in the data. Here is a before (top) and after (bottom).