Optimal Predictive Control for Maximum Utilization of Heterogeneous Battery Energy Storage System

The development and integration of renewable energy technologies into the bulk power system during the past few decades has been pushed by the disastrous effects of climate change, rising geopolitical tensions, and expanding energy needs. Furthermore, governmental agencies all around the world accelerated the decrease in cost of renewable energy-based generation and integration. However, a large percentage of intermittent energy production is dependent on solar radiation and wind speed which pose problems for the stability and resilience of the grid. Thus, upcoming power grid with high penetration of renewable energy resources requires widely accessible energy storage such as battery energy storage systems (BESS). However, BESS is expensive for large scale grid-integration and other demanding applications such as electric vehicles (EVs). High cost of mining the material associated with the battery and the high adoption of EVs may put more supply strain on these materials. The exponential increase in the number of EVs represents an opportunity to utilize the second-hand used batteries from EVs for power grid applications. This solution is not only cost affective but also save these materials from ending up in landfill or consuming more energy to be recycled while these batteries could be giving a second life. The main challenge with these second-hand used batteries are their heterogeneous characteristics and realization an effective power electronics interface and control scheme for their maximum utilization. Thus, this thesis proposes a control scheme for optimal selection of switching sequences in cascaded multi-level inverters (CMI) interfaced heterogeneous battery energy storage systems which are retired from EVs and need to be used in power grid application. The challenge is the sparse life of these battery cells and consequently different state of the charge (SOCs) in a string which limits their utilization. This thesis leverages the characteristics of CMIs to mitigate the impact of the sparse life of these battery cells via an optimal predictive control scheme for balancing the power drawn from CMI cells by taking into consideration the SOCs of battery cells. Several case studies are provided in this thesis that demonstrate significant improvement in utilization of heterogeneous BESS comparing to state-of-the-art control schemes.