A dependable, prognostics-incorporated, N-S modular battery reconfiguration scheme with an application to electric aircraft

Abstract

The design of dependable systems, such as electric aircraft, necessitates reliable battery system management to help ensure that in addition to power demands being met in the present, they can be met in the future with a low probability of failure. The development of rules concerning battery discharge can minimize the possibility of faults leading to failures due to extreme operating conditions such as exposure to excessive heat or deep discharge. Fault mitigation methods of battery system management can also include switching mechanisms that reconfigure the battery system to exclude faulty batteries. In our work, we propose a dynamic battery reconfiguration scheme for systems that contain battery packs. The battery reconfiguration scheme was designed with the motivation to be both dependable and capable of having a longer discharge time when compared to a static configuration. The contributions of the battery reconfiguration scheme are many. The scheme allows any battery to be bypassed and allows any primary battery to be functionally replaced by a spare battery resulting in fault tolerance. The scheme uses a minimal set of switches that is comparable to other battery reconfiguration schemes and uses diodes to facilitate safety by preventing batteries within the system from charging one another. The switching problem is solved in a general way using a constraint solver algorithm and the switches are controlled automatically by a microcontroller. The scheme facilitates maintainability as well as adaptability by using a modular design. As stated before, through use of prognostics and diagnostics, the discharge time of the battery reconfiguration scheme can be extended to longer than that of a standard, static configuration. The prognostic measure of remaining discharge time is used to assess the batteries that are limiting to the remaining operating time of the system. Since the limiting batteries can be subject to change, prognostics are taken at a regular interval. As the prognosis changes, the configuration is then switched to supplement the discharge of the limiting batteries or switched to allow the limiting batteries to rest which extends the remaining discharge time. The diagnostic measure used is the state of charge. We use both simulation of an electric aircraft battery bank and a magnetic levitation vehicle as a high current draw to show that the scheme is both fault tolerant and capable of extending the module discharge time. Our battery reconfiguration scheme and module design can be used with any vehicle that has similar power demands such as an electric aircraft. In essence, the key advantages of our battery reconfiguration scheme are that it is fault tolerant by allowing for the continued operation of the vehicle given a fault in a battery and that it can enhance the module discharge time which can be used either persistently to increase the operation time per cycle or upon invocation by flight plan contingency software in instances where the overall remaining discharge time in the current configuration is not satisfactory to complete a flight path.