Methods for State-Of-Function Algorithm Validation 2017-01-1219
Validation of the State-Of-Function (SOF) algorithm and associated cell models are critical for battery management as they are responsible for optimal pack power utilization as well as safety protection and life. The SOF accomplishes this optimization task by communicating pack level operation limits related to power, current, voltage and temperature. These operation limits are, in some cases, estimated via parameters and equations derived from cell models. Correspondingly, any errors within the cell models will propagate into the model-dependent SOF limits.
Understanding the source of errors and thus finding areas for improvement requires a visualization-based SOF validation strategy. In this work, we define and demonstrate an automated validation approach comprised of: (1) creating an array of pulse events that span the operation boundaries for the application of interest; (2) modeling the cell and SOF pack response for each pulse event (3) measuring the cell and SOF pack response for each pulse event (4) comparing the modeled and measured responses for each pulse event, and (5) creating an interactive software environment to explore each pulse event visually.
Our initial goal is to develop an environment in which model, SOF and measurement responses can be qualitatively contrasted and compared in order to facilitate rapid error identification and correction.