Optimal reconfiguration for vulnerable radial smart grids under uncertain operating conditions

https://doi.org/10.1016/j.compeleceng.2021.107310Get rights and content

Highlights

  • Efficient smart grid reconfiguration using Metaheuristic Manta Ray foraging optimizer.

  • Minimized power losses and improved voltage profile for reconfigurable radial distribution network under load variation.

  • Enhanced DN topology under contingency conditions for the IEEE 33-bus and 85-bus radial systems.

Abstract

Modern smart grid prospects necessitate handling abnormal operating conditions besides conventional demands for improving power systems capabilities. Uncertain load and generation, and line outages during contingency conditions of electric power systems should be properly and efficiently dealt with. Lately, lockdown situations because of COVID-19 pandemic have greatly influenced energy demands in many areas in the World. Vulnerable operation of power networks, especially in either isolated microgrids or large-scale smart grids can be significantly avoided through proposing optimal reconfigurable network. In this paper, employing the distribution network (DN) reconfiguration is deeply studied for achieving fault-tolerance and fast recovery to reliable configurable DN in smart grids. Since radiality is among crucial properties of DN topology, searching for feasible configuration of DN is NP-hard optimization problem. Therefore, the recent Manta Ray Foraging Optimization (MRFO) is considered for solving such DN optimization instance. Performance of MRFO is examined against two common optimizers: the Particle Swarm Optimization (PSO) and the Grey Wolf Optimization (GWO). Different operating conditions for both the IEEE 33-bus and IEEE 85-bus systems are analysed using these optimization techniques. The goal is to search for feasible reconfigured DN with the minimum power losses and the optimal enhanced voltage profile. Simulation results reveal that the proposed MRFO approach provides efficient and outstanding behaviour in various operation scenarios. The efficiency and the robustness of the proposed MRFO approach are verified, the power loss reduction ratio ranges between 21% and 41% in different studied scenarios and adequate voltage profile enhancement is achieved.

Introduction

Electrical power systems are susceptible to unexpected operating conditions. Uncertain time variant load and generation, the faults and the contingency operations of electric power systems should be properly and efficiently dealt with. Severe weather conditions such as storm and hurricane may impact the efficiency of distribution network (DN) for days or few weeks [1]. In such conditions, branch failure which causes line outage impacts in losses in different electrical infrastructure due to interruption in power electricity [2].

Recently, the clear impact of the exceptional circumstances during the coronavirus pandemic period on the worldwide generation and demand has been illustrated in the annual IEA Global Energy Review [3]. Monthly electricity demand reduction during full lockdown reached approximately 20% on average according to daily reports coming from more than 30 countries [3]. Moreover, expectations of 2020 refer to global electricity demand reduction down to about 5% and 10% in some regions. Variability of load and distributed generation (DG) can negatively affect load balancing and voltage index [4].

Keeping survivability of ad-hoc microgrids such as electric shipboard power requires high practical fault management system [5]. Reliability of distributed energy sources such as plug-in electric vehicles and wind farms have been conditioned to considering severe and contingency conditions as well as normal operation [6]. Real-time DN systems should overcome various situations as equipment maintenance, fault occurrence and out-of-area services [7]. Moreover, in dynamic DN systems, remotely controlled switches may affect the accommodation of DGs especially under uncertain load uncertainties [8]. Uncertain conditions in solar irradiation of solar-based generation, wind speed and unbalanced load have been addressed for multi-phase DN in [9]. Keeping distribution system reliability against high impedance faults has been studied in [10].

In this paper, DN reconfiguration is employed to ensure improved system characteristics in either normal or uncertain operating conditions. Three operation cases are studied and analysed, namely normal condition (called base case),load variation, and contingency line outage. The safe and reliable limits of both load active and reactive power expansion are analysed for IEEE 33-bus and 85-bus systems. The recent Manta Ray Foraging Optimization (MRFO) algorithm [11], under the proposed approach, showed outstanding performance when compared to both Particle Swarm Optimization (PSO) [12] and Grey Wolf Optimization (GWO) [13].

Moving towards smart grids technology became crucial for the sake of better management, fault detection and automatic DN recovery and reconfiguration especially during abnormal operating conditions [14]. Recent research studies in smart grids have been directed towards addressing new emerging cases other than achieving conventional demands particularly with the lack of full information of the nature of uncertainties in DN [15], [16]. Salau et al. [17] have considered varying both active (P) and reactive (Q) power loads for the IEEE 33-bus system. Their study has addressed applying DN reconfiguration in case of light load (0.5 × P, Q) and heavy load (1.3 × P, Q) conditions, other than normal operation. Devabalaji et al. [18] have applied load variation from light condition (0.5 × P, Q) to peak condition (1.6 × P, Q) with step value 0.01 for the IEEE 34 and 85-bus systems. The capacitors in the DN have been injected in order to raise its performance in case of mentioned scenarios. Essallah et al. [19] proposed placement of DG units with recalculated size and location for each different load level. Considered scenarios included only raising P up to 50% and increasing both P and Q up to 50% for the IEEE 33 and 69-bus systems. Wang et al. [4] have addressed the system stability in presence of different kind of loads such as residential/commercial/industrial. The DN reconfiguration has been employed to the IEEE 33, 69 and 118-bus systems, together with placement of DGs like PV panels, wind turbines and gas turbines for safe operation in case of daily load variability. Tyagi et al. [20] have considered non-statistical uncertainty conditions with ±5% around normal load and renewable power generation with ±20% of rated amount of generation under the IEEE 33-bus system.

In addition, the variation in power generation, in particular with renewable resources, has been considered in some studies in literature. For example, Cui et al. [21] have proposed a soft strategy for DN reconfiguration together with energy storage systems to handle side-effects of uncertain wind power. The strategy has been tested for the IEEE 16 and 33-bus and the PG & E-69 systems. Yong et al. [22] have studied active DN under possible uncertain conditions due to randomness of renewable DGs delivered power in the IEEE 33-bus system. Dinakara et al. [23] have examined the placement of DGs with renewable resources for the IEEE 15, 33, 69 and 85-bus systems.

For most of aforementioned critical operating conditions and possibly others, DN reconfiguration has been implemented as a tool for system recovery and survivability. DN reconfiguration is important for microgrid optimality, which is a step towards smart grids [24].

In the state of the art, DN reconfiguration has been commonly treated as an optimization problem. The graph-based model can properly represent the concerned DN. DN graph can be constructed using nodes (acting system buses) and edges (acting system branches or lines). The set of edges of DN graph represents the search space when DN reconfiguration is targeted. In this context, DN reconfiguration belongs to NP-hard (i.e, no polynomial-time solution) combinatorial optimization problems [7], [25]. Therefore, different soft computing approaches have been considered to tackle this problem. However, metaheuristic optimization algorithms are widely used and preferred, see for example [25] and many others. Flexible manipulation of metaheuristic techniques as well as high portability to integration and hybridization with other techniques are behind their successful application in DN reconfiguration problem., i.e.: genetic algorithms (GA) [26] and numerous following techniques. Regarding the study of uncertainties within DNs, PSO [12] may be among most frequently used optimizers. As well as GWO introduced in [13], it has been considered in various related studies, i.e.: design of microgrids with reconfigurable topology under severe uncertainty [15] and energy management of hybrid emergency power unit for aircraft application [27].

Although the rich literature review, searching for the most efficient optimization technique for reaching the optimal DN reconfiguration with the best voltage profile and minimum power losses under different operating conditions and uncertainties is still a technical challenge. Studying the effect of varying each of active power and reactive power loads individually and simultaneously, in particular, from the prospective of both respecting DN voltage constraints while loads elevation, and illustrating DN recovery in case of line contingency in different locations in DN topology are among the most interesting related research gaps. To attain this objective, current work presents a novel approach using the recent MRFO technique.

The proposed approach relies on implementing complete DN graph (i.e., with all initial TS in close status). Thus, preparing search space by extracting the loops. Then, the optimization process starts by collecting initial feasible population which positively impacts on speeding up the whole procedure by avoidance of infeasible solutions. Fitness function coding is taking into consideration integer solution encoding as well as voltage and current constraints. It is defined as single-based objective where total power losses minimization is requested to evaluate candidate solutions. Consequently, voltage profile of DN buses can be improved. Proposed approach using MRFO algorithm shows notable efficiency and robustness according to the analysis of the standard deviation and the success rate of obtained solutions through several independent computer runs.

The novelty of this research is mainly the application of such advanced MRFO algorithm under the proposed approach for DN reconfiguration optimality. The MRFO, introduced by Zhao et al. [11], has been inspired by the foraging behaviour of manta ray marine creatures. MRFO is a promising optimization technique which outperforms many other traditional ones for the known test functions. MRFO has a novel search mechanism through moving between different update procedures of found solutions which aims to strengthen the chance of avoiding local optima solutions and hence finding near-optimal solution. Such recent optimization technique has been successfully applied in different studies in power systems field, e.g.: optimal allocation and sizing of renewable DGs in DNs [27] and other studies.

The crucial contributions of the study are the consideration of:

(i) Studying load variation case with different scenarios considering different active and reactive power scenarios. Determining safe operating limits of load variation to maintain the DN stability and ensure its feasibility especially under critical circumstances during lockdown periods of COVID-19 pandemic.

(ii) Analysing contingency case for different located branches in DN and achieving both recovery from hazard condition and improving DN performance.

(iii) Employing the recent MRFO algorithm into DN reconfiguration to handle both normal and abnormal operating conditions of DN.

(iv) Verifying how efficient the proposed recent MRFO is compared to both PSO and GWO for: realizing system recovery from faults; reducing the total power losses and improving the voltage profile. To attain an impartial comparative analysis, the quality of the reported solutions via mean value, robustness and precision via standard deviation and success rate is achieved.

The rest of this paper is organized as: Section 2 illustrates the problem formulation where objectives and constraints are defined, as well as studied systems. Section 3 introduces the proposed approach, describing in detail the main steps. The optimization results for the 33-bus and 85-bus systems are given in Section 4, including an impartial technical comparison of the proposed MRFO optimization technique against other optimizers. The conclusions and future research directions are highlighted in Section 5.

Section snippets

Studied systems and problem formulation

In smart grids, fault detection, fault recovery and automatic network reconfiguration are so vital tasks [14]. Operational demands, rather than capital objectives, are considered in their planning process. The DN consists of many branches connecting the buses. Initially, it is considered that most of the branches have close status which are usually called Section Switches (SS). DN initial configuration is defined by naming the set of initial TS whereas buses are connected via SS. In the DN,

Proposed approach

For realizing an efficient, robust and practical optimization process for DN reconfiguration problem, the proposed approach summarized in Fig. 3 consists of three main steps. First, problem search space is prepared. Second, initial feasible population is collected. Third, metaheuristic optimizers are called to look for best solutions (i.e., new DN reconfigurations) according to defined fitness function under problem constraints.

Problem search space. Starting with topological data of buses, SS

Configuration

The proposed optimal DN reconfiguration is examined for System I: IEEE 33-bus shown in Fig. 1 and System II: IEEE 85-bus shown in Fig. 2. System I represents a small-scale DN with 5 initial TS. System II is considered as a large-scale DN with 8 initial TS. Locations of initial TS and values of resistance and reactance of lines can be found for both systems in Table 6, Table 7 in Appendix, respectively. According to the previous studies in literature [25], [28] and many others, the TSs data and

Conclusions and future trends

This study addresses employing power distribution network reconfiguration as a tool for optimizing the distribution network performance in either normal or abnormal conditions. The importance of the study stems from degradation effects because of uncertain load variations as well as critical circumstances on distribution networks. The recent lockdown during COVID-19 pandemic periods have led to global uncertain time-variant reduction of energy demand. In such line outage or contingency cases in

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

H.S. Ramadan obtained his Ph.D. from SUPELEC, France in March 2012. He was promoted as Associate Professor in Zagazig University, Egypt in 2017. In 2021, Dr. Ramadan joined the ISTHY in France. He was the guest and managing Editor in different Elsevier journals, author of 80+ high-ranked journal and conference papers. His main fields of interest are power systems control and optimization, renewable energy, hydrogen reservoirs, storage systems, and smart grids.

References (30)

  • MosaadM.I. et al.

    Power quality enhancement of grid-connected fuel cell using evolutionary computing techniques

    Int J Hydrogen Energy

    (2018)
  • ZhangD. et al.

    An improved TS algorithm for loss-minimum reconfiguration in large-scale distribution systems

    Electr Power Syst Res

    (2007)
  • PrakashD.B. et al.

    Optimal siting of capacitors in radial distribution network using whale optimization algorithm

    Alex Eng J

    (2017)
  • KhomamiM.S. et al.

    Bi-level network reconfiguration model to improve the resilience of distribution systems against extreme weather events

    IET Gener Transm Distrib

    (2019)
  • RajalakshmiN. et al.

    Performance enhancement of radial distributed system with distributed generators by reconfiguration using binary firefly algorithm

    J Inst Eng (India): Ser B

    (2015)
  • Cited by (0)

    H.S. Ramadan obtained his Ph.D. from SUPELEC, France in March 2012. He was promoted as Associate Professor in Zagazig University, Egypt in 2017. In 2021, Dr. Ramadan joined the ISTHY in France. He was the guest and managing Editor in different Elsevier journals, author of 80+ high-ranked journal and conference papers. His main fields of interest are power systems control and optimization, renewable energy, hydrogen reservoirs, storage systems, and smart grids.

    A.M. Helmi obtained his M.Sc. and Ph.D. in Computing from Computer Science Dept., Universitat Politècnica de Catalunya (UPC), Barcelona, Spain in 2010 and 2013, respectively. Since 2013, he is a lecturer at CSE Dept., Engineering Faculty, Zagazig University, Egypt. He is interested in research areas like human activity recognition, sensory data processing and applications of optimization techniques in engineering applications.

    This paper is for special section VSI-sgmg. Reviews processed and recommended for publication by Guest Editor Dr. H. H. Alhelou.

    View full text