Real-Time Distributed Control of Battery Energy Storage Systems for Security Constrained DC-OPF Security (IEEE 2017).


The introduction of fast-response battery energy storage system (BESS) provides a number of advantages to address the new challenges of smart grids, including improving the reliability and security of the power grids. This paper proposes a real-time distributed control algorithm to the corrective security-constrained DC optimal power flow problem for transmission network. The objective is to minimize the adjustment of the BESSs, while maintaining the supply-demand balance and ensuring no security constraint violations in the post-contingency state for the short-term period. Comparing to the conventional centralized methods, only simple local computation and information exchange with neighbors are required to update the local control signal, which leads to fast response of the BESSs to alleviate the impact of the transmission line outage. Real-time simulation results of the modified 6-bus and 24- bus systems demonstrate that the dynamic performances of proposed distributed algorithm satisfy standard requirements and indicate its applicability for practical power grid. Index Terms—Distributed control algorithm, real-time implementation, security-constrained, battery energy storage, DC optimal power flow.


THE ever growing demand, high penetration level of renewable generation, and the increasing complexity of power systems, pose new challenges to the power grid security issues [1, 2]. A large number of energy storage systems have already been installed to provide essential services such as frequency support, load shifting and shaving, intermittency handling for renewable generation, etc. Fast-response battery energy storage system (BESS) can be also utilized to improve the reliability and security of the power grid [3]. Contingency analysis as a routine is carried out in the operation of power systems with the purpose to ensure the system is balanced and reliable in both pre-contingency and post-contingency states [4]. N-1 security constrained optimal power flow (SC-OPF) refers to satisfying the security constraints when subjected to only one sudden component failure in the system. The existing models for SC-OPF problem can be classified in two major categories: the preventive security constrained model (PSC-OPF) [5] and the corrective security constrained model (CSC-OPF) [6]. The preventive model requires no re-adjustments of the control variables in all post-contingency states while satisfying all security constraints. Thus, the economic benefit in the pre-contingency state may be affected to a large extent, or even no feasible solution can be obtained in some situations for the preventive model. The feasibility of CSC-OPF is built upon the assumption that the remaining components are capable of withstanding the operational limits violations for a short period of time in post-contingency state, which buys some time for corrective actions to take effect [7].


Traditional CSC-OPF aims to minimize a predefined objective function in the normal operating condition, and guarantees the equality constraints (i.e. net power injection at each bus should be zero), and inequality constraints (i.e. generation output limits, transmission line limits should not be exceeded) are satisfied in both pre-contingency and post-contingency states. In addition, the re-adjustments of the control variables between precontingency state and each post-contingency state must be within the permissible limits. The conventional CSC-OPF problem can be formulated as follows [29].


In the proposed distributed algorithm, the Lagrange multipliers are discovered in a fully distributed way, and each local controller updates the control signal through information exchange with its neighboring buses connected by transmission lines. Moreover, the proposed distributed algorithm to CSCDCOPF problem guarantees the supply-demand balance and no violations of system constraints. The local controller updates the Lagrange multiplier through information exchange with its neighboring local controllers according to the following consensus based distributed algorithm:


To demonstrate the effectiveness of the proposed real-time distributed algorithm for CSC-DCOPF problem, simulation with a 6-bus system is implemented in Matlab/Simulink environment in Case study 1. Case study 2 investigates the performance of the proposed distributed algorithm on a modified 24-bus test system. A. Case study 1: 6-bus system As shown in Fig. 1, a 6-bus system is consisted of three generators, three loads and ten transmission lines. The step sizes for σ, τ, ρ, and ξ are set to 0.002, 0.005, 0.001, and 0.005, respectively, to make the radius of M less than 1 through trial and error. The discrete time interval of the control reference update for BESS is chosen to be 10 s, which corresponds to a balance between control performance and technical feasibility.


This paper proposes a fully distributed real-time algorithm to the CSC-DCOPF problem for transmission network, relying only on local computation and information exchange among neighboring controllers, which can distribute the computation burden and relief stress for the central controller. Comparing to the conventional centralized methods, the reduced amount of information to exchange contribute to reducing the cost of the supporting communication network. Moreover, the calculations carried out by local controller are decreased significantly, which contributes to fast response of the BESSs. The proposed distributed algorithm coordinates the available BESSs to respond in timely manner in the post-contingency state, in terms of maintaining the generation-demand balance and avoiding the violations of adjustment limits and load flow limits of the rest transmission lines. As demonstrated in the real-time simulation using Matlab/Simulink, the proposed distributed algorithm converges fast and the dynamic performances of the frequency and voltage magnitude satisfy standard requirements. Thus, the proposed distributed real-time algorithm has promising applications to the practical power grids.