To rank System Identification (SI) requests of multi-rail power converters based on system variation severity in real time, this paper presents a ranking approach that dynamically updates priorities in hard-real-time. By comparing quantitated application importance and severities of system variations of rails, their ranks can be acquired. Besides, a window time is set aside in SI processes, only during which higher priority requests may interrupt SI processes of lower priority rails so that estimation opportunities can be provided to both high and low priority rails in rush hours. The paper also proposed to save intermediate iteration results, acquired from interrupted SI processes, for being continuously used once the processes recover to save computational cost. For validation, this technique is applied for a five-rail power conversion architecture. The proposed workflow also shows the high flexibility as all variables included in this approach are user-defined to match individual requirements.
|Title of host publication||2021 56th International Universities Power Engineering Conference (UPEC)|
|Publication status||Published - 30 Sep 2021|
|Event||56th International Universities Power Engineering Conference - Middlesbrough, United Kingdom|
Duration: 31 Aug 2021 → 3 Sep 2021
|Conference||56th International Universities Power Engineering Conference|
|Period||31/08/21 → 3/09/21|