Abstract:
Electric aviation represents a new arena for battery engineering and development. In contrast to automotive applications, the electrification of aviation and aerospace is both less mature and requires higher safety and performance regulations. This work addresses a first step towards the development of standards and algorithms for measuring remaining useful energy for the battery system. Battery pack flight test data from 134 tests and two different manufacturers was analyzed to determine the weakest cell blocks in the pack, defined as cell blocks having the lowest voltage at the end of the test. It was found that the maximum initial voltage and voltage integral were two features with predictive power. Using the first five minutes of flight test data, accurate predictions were made ~85% of the time, in contract to the status quo where ~30 minutes of flight test data may be required. Sources of error and pathways to improve upon this result are discussed, such as improving data logging, adding additional logic to handle well-balanced packs, and engineering new voltage features with greater predictive power. It is possible to increase the accuracy beyond 95% this way. Future work will build on this to develop algorithms and standards for on-line fuel gauging for electric aircraft.