The objective is this project is to develop a HI-based failure curve for at least one distribution asset class, and potentially for more asset classes. Secondary objective is to research, develop and demonstrate approaches that can be used to translate health index results produced from asset condition assessment programs into a condition-based failure probability.
The health index formulations applied within this study contain degradation factors that will be influenced by the operating and environmental conditions as well as the maintenance practices that are conducted upon the asset. Each degradation factor is weighted based upon its individual contribution to the overall health index score such as those developed in leading CEATI health indexing studies. The final score is normalized on a numerical scale from 0 (very poor) to 100 (very good).
It is recommended that this study that a single distribution asset class be selected as a starting point, particularly one for which failure is heavily driven by condition, one for which condition and failure data is readily available, and one for which replacement and retirement activities are a significant concern for member utilities. Likely candidates include pad-mounted transformers or switchgear, however other assets may also be useful.
This project will utilize historical performance and reliability data along with associated health index data for those same assets in order to establish a correlation between failure probability and the health index value. Through this correlation, other outputs can be provided, including the typical time it takes for an asset to transition between the various grades of the Health Index, from Very Good to Very Poor.
The final output of this project will be a report that will show an condition based failure curve for one asset class, as well as will evaluate alternative estimation procedures, and select an optimal approach that utilities can apply in order to derive condition-based failure probability functions from given health index values.