A bi-directional dc/dc converter model is investigated for a notional Medium Voltage DC (MVDC) shipboard power system to improve energy flexibility and deal with peak energy demand in shipboard power system. Surplus energy in the MVDC system during light load condition can be captured by energy storages distributed in local load zones through the bi-directional dc/dc converters and then can be used during heavy load condition or black starting of the MVDC system. In this paper, the derivation process of the small-signal average models of the isolated-type bi-directional dc/dc converter is presented for controller design. This paper also presents the controller optimization process using intelligent optimal searching algorithm, Particle Swarm Optimization, for optimizing dynamic and steady-state control performance of a bi-directional dc/dc converter. The control performance of the proposed controller is evaluated using frequency-domain analysis and time-domain simulation of the large-scale notional MVDC shipboard power system using the Real-Time Digital Simulator.
Recently, direct current (dc) power distribution technology has been investigated in many studies as a promising candidate for future power systems [1], [2] and [3]. The major changes in power technologies have resulted from the development of power electronic technologies. While current dc power converters are not yet comparable to ac transformers in terms of efficiency and reliability, the additional functionalities and compactness may offset these disadvantages in some applications. Considering the fact that most power electronic loads need dc power for end use or dc power interfaces, dc distribution systems are considered advantageous for both low and medium power level distribution. Moreover, since power converters can execute multiple functions including active flow control and power quality improvement concurrently, they can provide flexibility in power system control and management.
This research is targeting dc power systems in electric ships, specifically a notional U.S. Navy Medium Voltage DC (MVDC) shipboard power system. Since shipboard power systems have limited resources in limited space, its control objectives are more rigorous and strategic than terrestrial power systems with regard to protections, restoration, reliability, survivability, and so on [4] and [5]. The purpose of the study is to develop a notional MVDC power system for next-generation integrated electric ship. Since the MVDC system contains a lot of components such as turbine generators, power converters, energy storages, and critical loads (radar, pulsed loads, and motor drives), the major task of the Electric Ship Research and Development Consortium (ESRDC) and FSU/CAPS is to implement the whole MVDC system into a real-time simulation model considering hardware-in-the-loop (HIL) test. Therefore, the final product of the whole project is the real-time simulation model in the RTDS (Real-Time Digital Simulator™) environment. From this point of view, the goal of the work presented in this paper is to augment the simulation model of a notional MVDC system, developed by the ESRDC as illustrated in Fig. 1, with realistic bi-directional dc/dc converter models. With the availability of such a model in a large scale simulation model, this research subsequently focuses on improving energy efficiency of the MVDC system using bidirectional dc/dc converters.The configuration of the developed bi-directional dc/dc converter model is depicted in Fig. 2, originally proposed by Wang et al. [6], because it has the following important features required by shipboard power systems: (1) galvanic isolation between two voltage levels through a high frequency transformer, (2) full-bridge converters on both sides for high power application, (3) current-fed converter on the low voltage side and voltage-fed converter on the high voltage side, and (4) active clamping circuit on the low voltage side for zero-voltage switching. Compared to other configurations such as those employing voltage-fed converters on both sides [7], this configuration can smooth the power transfer due to the inductor and allow the independent control of the two converters in each operating mode.This paper presents the controller design procedures of the bi-directional dc/dc converter. In Section 2, the small-signal average models of the converter for buck mode and boost mode are derived. The small-signal average models can provide useful tools for evaluating control performance such as pole locations and stability margins. Section 3 describes controller design process of the bi-directional dc/dc converter. An intelligent optimum search algorithm referred to as Particle Swarm Optimization (PSO) is applied to optimize the controllers. Finally, the control performance is verified via frequency-domain analysis and real-time simulations of the developed large-scale shipboard MVDC system in Section 4.
A bi-directional dc/dc converter model has been implemented to a large-scale real-time simulation model of a notional MVDC shipboard power system. The small-signal average models of a bi-directional dc/dc converter are derived for designing controller and evaluating the performance and stability of the power converters. The controller design for the bi-directional dc/dc converter is complicated because of non-minimum phase zero and uncertain operations of the critical loads in the MVDC system. The input voltages and currents are easily affected by the random operation of the loads and also may contain harmonic distortions. The disturbance in the voltages of the MVDC system can affect the control performance of the bi-directional dc/dc converter. This paper presents a systematic optimization method of the converter controller using Particle Swarm Optimization. The proposed method is suitable for the bi-directional power converters in the MVDC shipboard power systems because the proposed optimization method using PSO can be directly applied to the nonlinear model and also consider various operating points with no change in the optimization process. Hence, the proposed method can improve the performance and robustness of the controller. Real-time simulations with the whole MVDC shipboard power system model are provided to verify the control performance.
