Year
2010
Abstract
This paper describes our experiences using both real and simulated tank data to quantify the errors associated with imperfect marking of start and stop times of tank events such as shipments and receipts. Both forward and backward data processing is assessed, in modest time intervals, to recognize events. Event marking methods evaluated include differencing, multi-scale principal component analysis using wavelets, and segmented regression. All methods are evaluated on both raw and smoothed data, and several smoothing options are compared, including standard filters, hybrid filters, and local kernel smoothing. The two main performance measures are the estimated location of the start and stop times of key events, and the associated signal levels.