Relaxed Alternating Direction Method of Multipliers for Hedging Communication Packet Loss in Integrated Electrical and Heating System

Abstract

Integrated electrical and heating systems (IEHSs) are promising for increasing the flexibility of power systems by exploiting the heat energy storage of pipelines. With the recent development of advanced communication technology, distributed optimization is employed in the coordination of IEHSs to meet the practical requirement of information privacy between different system operators. Existing studies on distributed optimization algorithms for IEHSs have seldom addressed packet loss during the process of information exchange. In this paper, a distributed paradigm is proposed for coordinating the operation of an IEHS considering communication packet loss. The relaxed alternating direction method of multipliers (R-ADMM) is derived by applying Peaceman-Rachford splitting to the Lagrangian dual of the primal problem. The proposed method is tested using several test systems in a lossy communication and transmission environment. Simulation results indicate the effectiveness and robustness of the proposed R-ADMM algorithm.

Scaled form Alternating Direction Method of Multipliers and its variants for solving the matrix equation AXB+CXD=E

“…The ADMM algorithm has numerous versions [24,[26][27][28]. We will briefly introduce two variants here (Dynamic Parameters and Relaxation).…”

“…Scholars in the field of engineering have paid attention to the problem. For example, Liang et al [24] proposed a relaxation scaled ADMM for coordinating the operation of integrated electrical and heating systems considering the loss of communication packets. There are also techniques for adjusting the quadratic penalty parameters [26][27][28].…”

Assessing the Impacts of Nonideal Communications on Distributed Optimal Power Flow Algorithms

“…Communications between the agents may occasionally fail entirely for multiple iterations. For instance, communication transmission collisions or instability of the communication link can cause packet loss preventing the agents from sharing data [16], [17]. To model communication loss, we consider an exponential failure distribution with a constant failure rate.…”

“…The impact of packet loss on unconstrained convex distributed optimization is investigated in [16]. A relaxed ADMM algorithm is proposed in [17] to optimize integrated electrical and heating systems considering communication packet loss.…”

Accelerated convergence of time‐splitting algorithm by relaxation method

“…Relaxed ADMM algorithm was used in refs. [30–32], in which the range of relaxation factor is $$(0, (1+\sqrt {5})/2)$$. The convergence proof of R‐ADMM can be found in ref.…”