Influence of load alterations to optimal network configuration for loss reduction
Highlights
► A novel heuristic algorithm is proposed aiming to simulate network reconfigurations in real time. ► Numerous scenarios have been examined in order to identify the sensitivity of the appropriate switching operations. ► The results confirm that load variations slightly affect the optimal solution even for extreme load variations. ► The algorithm results also provide the participating percentages in the reconfigurations of all sectionalizing switches. ► The results indicate that divergences from the actual optimal solution could be considered negligible.
Introduction
Loss reduction in power systems has constituted one of the most important objectives for researchers and engineers. The constant growth of energy demand along with the polluting conventional power plants has forced engineers in searching methods to reduce losses in all three stages of a power systems’ operation; generation, transmission and distribution. It is estimated that the largest proportion of losses in power networks corresponds to distribution networks; for a typical system in a developing country, distribution losses account approximately 13% of the total energy produced [1].
Over the past three decades considerable research has been conducted for loss minimization in the area of distribution systems. The basic concept for loss reduction, developed by Merlin and Back [2], aimed to take advantage of the distribution networks’ structure. Although distribution systems are designed as meshed networks, they operate as radial ones due to reliability and short circuit issues. The existence of tie switches that interconnect feeders and permit load transfer among them has lead to the idea of network reconfiguration for loss reduction. Changes of the network topology are performed by opening sectionalizing (normally closed) switches and closing tie (normally open) switches. All the needed switching operations are implemented in such a way that a number of constraints, i.e. voltage and current limits, radial structure of the network, etc., are not violated.
Network reconfiguration for loss reduction has been treated by many researchers and through a great number of different approaches. Although Merlin and Back [2] were the first who introduced the concept of distribution system reconfiguration (DSR), Civanlar et al. [3] proposed a purely heuristic algorithm based on a branch exchange method. By this approach they proposed an approximate formula in order to estimate whether a particular switching operation would increase or reduce losses. Shirmohammadi and Hong [4] based their algorithm on the approach of Merlin and Back including optimal power flow as the basic criterion for the switches that should open. Baran and Wu [5] attempted to improve Civanlars’ method by introducing two approximation formulas for power flow. Moreover, in [6] the methods concerning loss minimization algorithms published in IEEE transactions between years 1988 and 2002 are presented. The reconfiguration algorithms may be classified by their solution methods in three basic categories; mathematical optimization methods, heuristics, and those based on Artificial Intelligence. In [7] the authors present a mathematical model for loss minimization which consists in introducing non-conventional group of variables instead of the classical bus complex voltages. The main idea is to simplify the mathematical optimization problem by eliminating continuous and binary variables. The result is to formalize the minimization problem with a linear objective function. Heuristics have kept being proposed by researchers for loss minimization due to their simplicity. In [8] a heuristic algorithm is proposed based on the direction of the branch power flows while in [9] the reconfiguration problem is solved by a heuristic approach and rules base. The proposed simple rules are formed based on the system operation experiences which in turn enhances the heuristic nature of the algorithm. Some more sophisticated approaches involve artificial intelligence techniques like the genetic algorithms presented in [10], [11]. More specific the algorithm presented in [10] is actually a meta-heuristic searching algorithm that combines high local search efficiency with global search ability of intelligent algorithms. The latter constitutes the basic idea presented in [12] where the proposed algorithm is based on a fuzzy approach with some heuristic rules.
In all aforementioned papers loss reduction by network reconfiguration is treated for fixed operational points, assuming constant load demand for all nodes of the examined topologies. In most of these cases, load demand is considered as the peak value. Although this practice offers a common base for the evaluation of the efficiency of all proposed algorithms, it is not suitable for simulation of real operating conditions. In practice, load patterns indicate load variations concerning the networks’ consumers, which can fluctuate in high levels for different customer types. Therefore, it becomes obvious that optimal reconfiguration should adapt to account for load variability, in such way, that the frequency of the reconfigurations coincides with the assumed time periods for which minimization of losses is desired.
Broadwater et al. [13] were one of the first teams that tried to incorporate load variability in the reconfiguration problem. A simple case study was used to illustrate that when different load patterns are applied, optimal reconfiguration can actually alter subject to the aforementioned different load conditions. Moreover, in [14] the concept of short and long term operation of distribution systems is introduced. The above approach aims to simulate actual load conditions. An hourly optimal switching algorithm is utilized for the determination of the hourly optimal configuration, whereas concerning the long term operation a method is adopted for the seasonal operation of the network. Peponis et al. [15] focused their analysis in load modelling for the purposes of network reconfiguration, but they also examined load variation with respect to optimizing reconfiguration decisions. In [16] an on-line approach for loss reduction is presented based on artificial intelligence. The proposed method is based on learning classifier systems which continually propose configurations in the case of time-varying profiles of energy requirement. Huang and Chin [17] also applied actual load patterns for different customer types in their examined topology in order to identify the switches that had to change their status during specific time periods of the day. In [18] an hourly reconfiguration is evaluated compared to fixed topologies, considering maximum and average demand of the system. The paper concludes that hourly reconfiguration seems not to be so effective as compared to a simple maximum or average demand configuration. Bueno et al. [19] examined in their work a typical 24-h period for low, medium and high load values. In this case, they concluded that although the optimum loss reduction is achieved when network configurations are altered to adapt to load variations, however, an optimal fixed configuration for a specific load level and time period does not necessarily lead to a significant increase of losses.
In this paper, network reconfiguration for loss reduction has been originally obtained for a fixed operational point. Load variations, considered to simulate actual load conditions, are taken afterwards into account in order to investigate if the previous configuration is modified. The methodology adopted in this work has been applied to some of the most common used distribution topologies in the published literature, i.e. 16, 33 and 69 bus systems, as well as to a real urban distribution network of the city of Thessaloniki, Greece. In this work, stochastic active power of each examined networks’ nodes is assumed to be a stochastic variable following uniform distribution.
This paper is organized as follows. In Section 2, the problem formulation is illustrated along with aspects concerning the time-varying loads. In Section 3, the algorithm developed for loss reduction taking into account the load variability is presented. In Section 4, case studies along with their specific parameters are shown. In Section 5 the results of the simulations are presented and finally in Section 6 the conclusions derived are discussed.
Section snippets
Fixed operational point
Active power electrical losses in power systems are proportional to the square of branch current. The problem of loss minimization in distribution networks can be written for a fixed operational point in a simple form as follows [18], [20], [21], [22]:subject to:where n is the total number of nodes, m is the total number of branches, ns is the number of sources, I is the m-vector complex branch current with, C is the n-vector
Proposed algorithm for network reconfiguration
The algorithm proposed in this paper for network reconfigurations is based on a heuristic approach which in turn adopts two basic rules commonly used in most of the published heuristics algorithms, such as [5], [6], [7], [8], [9], [10], [11], [12], [13], [14], [15], [16], [17], [18], [19], [20], [21], [22], [23]. In the last years many papers with new heuristic algorithms applied to the problem were published. In [24] an effective two-stage method for DSR for loss minimization is presented,
Examined networks
The algorithm described in the previous section was applied on the 16, 33 and 69 bus systems. In Table 1 basic data for the networks are presented.
A typical urban power distribution network segment was selected as the real test case in the study. The selected segment lies in the eastern part of Thessaloniki, Greece. It consists of five MV power lines which start at the same HV (high voltage)/MV power substation, and run mostly underground, with only a small section consisting of overhead lines.
16 bus system
In Fig. 4 the layout of the IEEE 16 bus system is presented. It was expected that due to the small network size, load alterations would not affect significantly the switching operations for optimal configuration. The proposed algorithm was applied for the two deviation values and for 10,000 different loading scenarios.
The results concerning the solution g0 show that the close-open switching operations correspond to pairs Tie-1/branch 8–6 and Tie-2/branch 5–7, shown in Table 2. Operating switch
Conclusions and discussion
The first contribution of this paper is an expansion of network reconfiguration for loss reduction in a way to incorporate load alterations and thus simulate real operating conditions. The conclusions derived by such approach may constitute the guideline for centralized real management of the network for loss reduction. Secondly, the proposed algorithm may be used as a forecasting technique for network reconfigurations considering that load conditions tend to follow normally expected annual
Acknowledgement
The authors would like to thank Messrs. S. Fahouridis and A. Zafirakis of the PPC for providing the needed data that enabled this study to be conducted successfully.
References (34)
- et al.
A survey of the state of the art in distribution system reconfiguration for system loss reduction
Electr. Power Syst. Res.
(1994) - et al.
A simpler and exact mathematical model for the computation of the minimal power losses tree
Electr. Power Syst. Res.
(2010) - et al.
A new heuristic approach for distribution systems loss reduction
Electr. Power Syst. Res.
(2008) - et al.
A rule based comprehensive approach for reconfiguration of electrical distribution network
Electr. Power Syst. Res.
(2009) - et al.
An improved TS algorithm for loss-minimum reconfiguration in large-scale distribution systems
Electr. Power Syst. Res.
(2007) - et al.
Distribution network reconfiguration for loss reduction by ant colony search algorithm
Electr. Power Syst. Res.
(2005) Reconfiguration of distribution system using fuzzy multi-objective approach
Electr. Power Syst. Res.
(2006)- et al.
Distribution feeder energy conservation by using heuristics fuzzy approach
Int. J. Electr. Power Energy Syst.
(2002) - et al.
A heuristic method for feeder reconfiguration and service restoration in distribution networks
Int. J. Electr. Power Energy Syst.
(2009) Optimal reconfiguration of electrical distribution network using the refined genetic algorithm
Electr. Power Syst. Res.
(2002)
Probabilistic power flow with non-conforming electric loads
Int. J. Electr. Power Energy Syst.
A geometrical approach for network reconfiguration based loss minimization in distribution systems
Int. J. Electr. Power Energy Syst.
Cost/worth assessment of reliability improvement in distribution networks by means of artificial intelligence
Int. J. Electr. Power Energy Syst.
Search for a minimal-loss operating spanning tree configuration in an urban power distribution system
Distribution feeder reconfiguration for loss reduction
IEEE Trans. Power Deliv.
Reconfiguration of electric distribution networks for resistive line loss reduction
IEEE Trans. Power Deliv.
Network reconfiguration in distribution systems for loss reduction and load balancing
IEEE Trans. Power Deliv.
Cited by (37)
Dynamic reconfiguration of distribution network based on dynamic optimal period division and multi-group flight slime mould algorithm
2022, Electric Power Systems ResearchEnergy System Planning towards Renewable Power System: Energy Matrix Change in Cuba by 2030
2018, IFAC-PapersOnLineHarmonic state estimation for distribution networks using phasor measurement units
2017, Electric Power Systems ResearchEffects of demand response programs on distribution system operation
2016, International Journal of Electrical Power and Energy SystemsCitation Excerpt :Network usage and operation efficiency optimization is the main goal of smart grid initiatives. Distribution network reconfiguration following hourly load variations for minimizing power losses is one of these initiatives [13]. Several proposals have been reported in the literature about distribution reconfiguration methodologies, which have the main goal of reducing network electrical losses by opening or closing switches to change the distribution electrical system configuration and keeping in mind voltage and power flow constraints.
Novel methodology for optimal reconfiguration of distribution networks with distributed energy resources
2015, Electric Power Systems ResearchCitation Excerpt :Feature 2—Addressing the DER uncertainty: In literature, this is handled in different ways. For example, the load uncertainty is handled by segregating the load profiles based on customer type [4], maximum and average demand of the system [27] or other stochastic procedures as in [14]. This paper uses SA method to account for DER uncertainty.