Year
2013
Abstract
The international safeguards community is interested in using UF6 cylinder weight data from process scales at feed and withdrawal areas to help verify operations as declared. Regardless of the data collection method and frequency, the raw weight data must be analyzed and presented in a way that will improve inspection efficiency. Oak Ridge National Laboratory (ORNL) is using a mock feed and withdrawal system to generate load cell weight data representative of actual operations. This system pumps water, used as a surrogate for UF6, from feed containers through a “process” and into tails and product containers. The weight of each container is monitored by a load cell weighing system. ORNL is developing a methodology to automatically analyze raw weight data to identify the number of cycles processed and the material transferred. Modular and scalable algorithms were developed by leveraging commercial software packages. The algorithms use a set of rules to represent normal operational states (i.e., vacant scale, full container, empty container, draining/filling container). Each datum is placed into one of these states by testing it against the rule set. The algorithms then identify complete cycles by comparing this sequence of states to a state diagram. This approach allows the rules to be abstracted from the algorithms. Therefore, the same algorithms can be applied to data collected from feed and withdrawal stations that perform differently.