Tank Measurement Data Compression for Solution Monitoring

A tank measurement data compression scheme is described that has been developed to meet the peculiar needs of nuclear materials safeguards in general, and solution monitoring in particular. The scheme has two stages, a filtering stage followed by a correlate and test stage. The filter might be viewed as an extension of the box car algorithm that enables feature extraction; the correlate substage uses a recursive least squares estimator, and the test stage is based on the Cusum test. The scheme has been developed by testing with data obtained from a commercially operated facility.