**1. Introduction**

Solar energy installations have experienced rapid growth in the last decade [1], which brings both environmental and economic benefits. One of the driving forces behind this growth is that the solar industry keeps seeing innovation for both reducing the system cost and increasing the grid-integration performance [2]. The mainstream DC voltage has been increasing from 600 V or 1000 V (low voltage in relevant standards [3,4]) to 1500 V, which is the maximum voltage of the low voltage directive according to the IEC standards in order to reduce the cost of large-scale photovoltaic (PV) systems. By doing so, for a given capacity, the installation cost can be reduced to a large extent (fewer strings, connections, and less cabling) [5]. On the other hand, the high variability of the solar PV energy (due to the weather conditions) raises challenges for integrating these PV systems to the grid. In such a case, energy storage could be integrated to the PV systems for smoothing the output power of PV plants. Recently, this has also become a promising solution toward smart PV systems [6].

For 1500 V PV applications, several studies have shown that single-stage conversion PV systems (without DC-DC stage) outperform two-stage based PV systems with respect to size, efficiency, and cost-effectiveness [7–9]. For the same reason, the battery energy storage systems (BESS) for large-scale PV Plants are also based on single-stage conversion [10,11], i.e., one bi-directional DC-DC or DC-AC stage depending on the type of connection: DC-coupling or AC coupling. As illustrated in Figure 1a, in DC-coupling, the output of the BESS is connected to the DC side of the PV inverter, while, in AC-coupling, as shown in Figure 1b, the BESS is added to the PV system at its AC side. Both of the configurations have the potential to improve the grid-integration performance of 1500 V PV systems. Flexible power managemen<sup>t</sup> is significantly enhanced in such systems.

**Figure 1.** System diagram of the single-stage 1500 V PV system with integrated battery energy storage systems (LF: low-frequency transformer): (**a**) DC-coupled configuration and (**b**) AC-coupled configuration.

Many research efforts have been devoted to address the design and control of PV-battery systems. In [12,13], the common methods for power smoothing and ramp-rate reduction with BESS were compared in terms of power tracking performance and BESS capacity requirement, respectively. In [14], the coordinated control of a single-stage based PV system with a DC-coupled BESS was analyzed along with energy management. The methods to determine the optimal sizing of BESSs were developed in [15], where second-life Li-ion batteries were considered for a cost-effectiveness analysis. The benefits of the DC-coupled BESS for a large-scale PV plant were investigated in [16], which shows a higher efficiency and less energy losses (with oversized PV arrays) when compared to an AC-coupled configuration can be achieved. Moreover, the benefits of the AC coupling over DC coupling were investigated in [17], in terms of reduced integration challenges and increased design flexibility.

However, the prior-art studies did not fully cover the discussions with respect to the lifetime and reliability of the PV-battery systems. As the loading on the converters in the two configurations is different, the reliability performance will also vary, which, in turn, may affect the final design. When considering that, recently, the reliability analysis of the power converters in PV-battery systems has attracted increasing interest, such as those presented in [18,19]. More specifically, a single-phase DC-coupled PV-battery system was considered in [18], where the impact of different self-consumption control strategies on the PV inverter reliability was analyzed. This work was further strengthened in [19] by considering the reliability analysis of the remaining power converters (i.e., PV boost converter and battery converter), in which the system-level reliability was also investigated when considering both DC-coupled and AC-coupled BESS configurations. However, the above investigation was performed for residential PV systems. When it comes to large-scale PV applications, e.g., 1500 V PV-battery systems, different insights in reliability analysis may be offered, which, in turn, can provide further design considerations to enhance the system performance. Nevertheless, such an analysis has not been thoroughly and systematically discussed in the literature. Thus, it is necessary to explore the power converter reliability for large-scale PV-battery systems, following which, a proper design of these power converters can possibly be achieved.

With the above concerns, this paper investigates the BESS of DC- and AC-coupling for 1500 V PV systems with emphasis on the reliability comparison. The analysis is carried out through a case study instead of a new methodology preposition on a 160 kW/1500 V PV system, in which proper BESSs are designed for the DC- and AC-coupled configurations to smooth the PV power and limit the power ramp rate. The rest of this paper is organized, as follows. In Section 2, the system modeling is presented, which includes the mission profiles and basic components for the system under study. In addition, the general control to enable the power smoothing and limit the power ramp rate is briefed. Subsequently, in Section 3, the reliability analysis is presented. First, a component-level reliability performance of all the power devices (i.e., IGBTs and diodes) within the two configurations is evaluated by estimating their lifetime under a real mission profile. Afterwards, the converter- and system-level reliability assessments based on the reliability block diagrams along with the Monte-Carlo simulations are carried out. Through this comparative reliability analysis, the most fragile part within the two systems can be identified; additionally, the overall system reliability can be improved by selecting an adequate BESS connection type, which is shown in Section 4. Following, the reliability benchmarking results are discussed further in Section 5. Finally, concluding remarks are given in Section 6.
