Jiliang Zhang, Tesla Motors (JZhang@teslamotors.com) Carolyn Wozniak, Tesla Motors (firstname.lastname@example.org) ABSTRACT Electrical Vehicle (EV) super charging stations provide electrical energy to incoming vehicles that need to be charged. Component failures may lead to a complete loss or reduction of power available to the vehicles. The reduced power still enables vehicle charging, but the charging time will be extended, which will require the customers in the queue to wait longer. On the other hand, the incoming vehicle flow demands a specific reliability requirement of the super charge station; therefore, providing the highest possible power at any time with minimum cost possible is in the interest of product design and service. Both system reliability capability and the behavior of the incoming vehicles play critical roles in system reliability modeling and analysis. In this paper, the multistate coherent structure (MCS) is employed to model the system reliability and availability of a single server (the charge port in our super charge application) and the system (charge cabinet with multiple charge ports or super charge station with multiple cabinets). The system state is defined as the available utility (energy output). The number of customers in service is assumed fixed or random with a probability distribution. The system reliability, availability, and component importance are defined accordingly. Two types of customers are considered: (1) a “stubborn” customer who will stick with a randomly selected server (charge port) regardless of the percentage of utility (charging power) available and (2) a “smarter” customer who can always choose the server with the maximum utility (charging power) available. Furthermore, customer waiting time in the queue is also considered. The service time is no longer statistically independent and identically distributed as assumed in queueing theory due to the fact that the partially failed server (cabinet) still provides a reduced utility (energy output), but the service time (charging time) will be extended. The performance measures are defined as (1) mean waiting time, mean service time, and mean sojourn time, and (2) mean queue length. The model and analysis method are applied to a simplified EV super charge station as an application example. A simulation tool is developed for system reliability, availability, and queue performance measures. Two design alternatives are analyzed and compared against each other. The design and service improvements are further discussed and proposed. The methodology presented here has been proven to be very useful in system reliability design optimization and field reliability improvement. Authors believe it will find more applications in super charge station design and service application and beyond.