Elsevier

Renewable Energy

Volume 102, Part B, March 2017, Pages 316-325
Renewable Energy

Distribution networks' energy losses versus hosting capacity of wind power in the presence of demand flexibility

https://doi.org/10.1016/j.renene.2016.10.051Get rights and content

Highlights

  • Multi-objective optimization model proposed for maximizing hosting capacity and minimizing losses.

  • The role of demand response option is investigated.

  • Multi-period OPF is used for HC maximization.

  • The benefits of DG owners as well as the DNOs are considered simultaneously.

Abstract

With the increasing share of renewable energy sources (RES) in demand supply, the distribution network operators (DNOs) are facing with new challenges. In one hand, it is desirable to increase the ability of the network in absorbing more renewable power generation units (or increasing the hosting capacity (HC)). On the other hand, power injection to the distribution network by renewable resources may increase the active power losses (if not properly allocated) which reduces the efficiency of the network. Thus, the DNO should make a balance between these two incommensurate objective functions. The Demand Response (DR) in context of smart grids can be used by DNO to facilitate this action. This paper provides an approach in which a multi-objective and multi-period NLP optimization model is formulated where the DR is utilized as an effective tool to increase HC and decrease the energy losses simultaneously. In order to quantify the benefits of the proposed method, it is applied on a 69-bus distribution network. The numerical results substantiate that the proposed approach gives optimal locations and capacity of RES, as well as minimum energy losses by load shifting capability provided via DR programs.

Introduction

The capacity of a distribution network for acceptance of new distributed generation (DG) units is called “hosting capacity” (HC). Calculation of the HC is an effective tool for determination of the most suitable locations for installation of DGs and hence, the investments will be guided toward the critical and most effective nodes of the grid. The distribution network operator/owner (DNO) is willing to increase this capacity since it would increase its benefits accrued due to connection fees. The presence of DG units has potential positive and negative aspects for DNO. The positive side of DG units include postponing the need for network reinforcement, reducing environmental pollution, active power loss reduction, network restorations [1] and voltage profile improvement. These potential benefits would become actual if the DG units as well as the distribution network are operated and planned optimally. But, there are some barriers which may hinder this transition. Those are namely as follows:

  • The intermittent renewable power generation: the main problem with renewable power generation resources is the volatility of their output [2] which is a function of different environmental parameters like temperature, solar radiation, wind speed [3] and etc. The only available option for controlling the wind power resources is its reactive power control or curtailment of its active power generation.

  • Different objective functions of DNO and DG owners: the DNO and DG owner/operator (DGO) do not have the same objective functions (or strategies). This would result in conflict of interests or a financial gain or loss. One simple explanation is that DNO is interested to attract more wind power generation capacity in its territory (the same as DGO). This is because the DGO can sell more energy and DNO can receive more connection fees. However this may increase the active energy losses which is not desirable for DNO.

The smart grid technology provides more flexibility for DNO to optimize its goals. One of these powerful flexibilities is called the Demand Response (DR). The DR describes an interaction and responsiveness of the consumers and offers different potential benefits in operating and planning of power systems [4].

This paper, models the concerns of both DNO and DGO, by considering a multi-objective framework. Thus, focus of this work is to maximize the HC in distribution networks, as well as to minimize the energy losses. A mid-term perspective (one year) is considered which explores how a distribution network as well as the wind turbines should be operated in order to attract more wind power generation capacity without deteriorating the network efficiency (i.e. without increasing the total energy losses). The DR in form of flexible (shift-able) demand is utilized along with other control options such as wind power generation curtailment and/or reactive power outputs of wind turbines. Any node which has non-zero demand can be used as the demand response node. No specific requirement is considered for the nodes to serve as demand response flexibility provider [5]. A technique for optimal DR node selection is provided in Ref. [6].

There are some factors which limit the HC including: voltage step limits, thermal loading limits of feeders and voltage limits. The previous works in this area can be generally categorized into two groups: the first group tries to identify the HC of a given network while the second group tries to maximize it. In Ref. [7], a hybrid GA-OPF is proposed to identify the best size and location for a predetermined number of thermal DG units. It is assumed that all DG units operate at constant power factor. This model uses a single snapshot of demand in different nodes of the network. To consider the intrinsic variability of electric demand, a probabilistic load flow using Monte Carlo technique was proposed in Ref. [8] that analyzes the role of power factor capabilities of PV systems on increasing the HC in distribution networks. The impact of harmonic distortions on limiting the HC is analyzed in Ref. [9]. The available models proposed for maximizing the HC use different techniques to do so such as energy storage units [10], network reconfiguration [11], OLTC and SVC control [12], curtailment of renewable energy resources [13], cost-benefit analysis [14], active/reactive power control of PV inverter [15], active network management [16] and grid reinforcement. Most of the researches are based on the optimal power flow (OPF) models. These approaches are proposed to determine the available HC for DGs [17] and wind power [18] in a distribution networks. Genetic algorithm (GA) [19], and hybrid OPF and GA [7] are utilized to determine the optimal position and size of DGs. In Ref. [20], multi-period OPF is used to determine the maximum wind energy HC of distribution networks. Moreover active management strategies such as wind power curtailment [18] [20], reactive power compensation, voltage control using coordinated on-load tap changer [18] [20], and power factor (PF) control of wind turbines [20] are investigated to increase the DG or wind power HC of distribution networks. In Ref. [17] authors proposed a methodology for determination of maximum DG capacity in radial low-voltage feeders. The methodology indicates the highest capacity that can be installed at a fixed point in the feeder for which the voltage is kept within the permissible limits in critical scenarios, especially in low load- high wind scenarios. In Ref. [18] a decentralized voltage control approach aimed to allow DG active power production maximization and to avoid DG disconnection due to voltage limit violation as much as possible. A local active/reactive power management control strategy was proposed based on Artificial Neural Networks, able to regulate voltage profiles at buses where DGs are connected, taking into account their capability curve constraints. In Ref. [19] probabilistic approaches were proposed to in order to determine maximum DG penetration in medium voltage distribution networks. In Ref. [20] curtailment is used to allow more wind or solar power to be connected to a distribution network when over-current or over-voltage occurs. In this regard, the concepts of “hard curtailment” and “soft curtailment” were introduced. A model based on cost benefit analysis is proposed in Ref. [14] for determining the optimal wind power HC of a distribution system using active-management strategies (AMSs). In Ref. [21] the economical benefits of different autonomous inverter control strategies for increasing the HC of a real low voltage grid in Germany. The costs of these strategies are compared with those of two alternative approaches, traditional grid reinforcement and a distribution transformer with OLTC. In Ref. [22], a probabilistic methodology is presented which integrates DR in real-time distribution energy market. The model proposed in Ref. [22], is day ahead and a single objective approach.

To the best of our knowledge, no work in the literature considers DR to increase the HC of distribution networks as well as minimizing the total energy losses. Accordingly, the contributions of current research are as follows:

  • 1)

    To investigate the impact of demand flexibility (in the context of DR program) on the maximization of wind power HC and minimization of energy losses in power distribution systems.

  • 2)

    The proposed model considers the contradicting objectives of DNO and DGO simultaneously.

  • 3)

    A multi-period AC power flow model is proposed to capture the variation of wind power and demand levels. All technical network constraints such as thermal rating of feeders and voltage limits have been considered, which makes the proposed model practical and realistic.

In this work, the demand level as well as wind power generations in different sites are assumed to be known. However these values are subject to uncertainty. There are several methods for handling the uncertainties existing in the defined problem such as stochastic methods [23], robust optimization [24] and Information gap decision theory [25].

The remainder of the paper is organized as follows. Section 2 describes the problem formulation and presents the modeling features and assumptions made in the proposed decision making framework. Simulation results and discussions are presented in Section 3. Section 4 concludes the paper.

Section snippets

Assumptions

  • The proposed model is run on a yearly basis (i.e. 56 periods with different durations are considered).

  • The DNO has the authority for controlling demands in some specific nodes (in the context of DR program). This can happen using mutual agreement/contract between the consumers and the DNO [26]. The gained benefits of this agreement will be shared between the DNO and the consumers.

  • The main idea of the proposed framework is to demonstrate and quantify the effectiveness of the developed model for

Data

The proposed algorithm is implemented in GAMS [27] environment and solved by SNOPT solver [28], running on an Intel®Xeon™CPU E5-1620 3.6 GHz PC with 8 GB RAM. It is applied to a 69-bus distribution network [29]. The proposed framework considers one year as the planning horizon. The total number of hours in each year are 8760 h. In order to reduce the computation burden, these 8760 h have been categorized into 56 clusters as given in Table 1 [11]. The duration of each cluster as well as the

Conclusions

In this paper, a multi-objective DR framework is presented in which the loss minimization as well as the HC maximization are analyzed. The simulation results show that the proposed strategy can be used by DNOs in practical cases. As evidenced by the simulation results, the proposed method offers some interesting features over traditional methods as follows:

  • It models the variation patterns of demand and wind power generation.

  • It can be utilized to assess the merits of nodes for participating in

Acknowledgements

The work of A. Soroudi was conducted in the Electricity Research Centre, University College Dublin, Ireland, which is supported by the Commission for Energy Regulation, Bord Gáis Energy, Bord na Móna Energy, Cylon Controls, EirGrid, Electric Ireland, EPRI, ESB International, ESB Networks, Gaelectric, Intel, SSE Renewables, UTRC and Viridian Power & Energy. A. Soroudi is funded through Science Foundation Ireland SEES Cluster under grant number SFI/09/SRC/E1780.

References (32)

  • S. Nojavan et al.

    Robust optimization based price-taker retailer bidding strategy under pool market price uncertainty

    Int. J. Electr. Power & Energy Syst.

    (2015)
  • M. Moradi-Dalvand et al.

    Self-scheduling of a wind producer based on information gap decision theory

    Energy

    (2015)
  • X. He et al.

    How to engage consumers in demand response: a contract perspective

    Util. Policy

    (2013)
  • M. Laumanns et al.

    An efficient, adaptive parameter variation scheme for metaheuristics based on the epsilon-constraint method

    Eur. J. Oper. Res.

    (2006)
  • B. Ansari et al.

    Restoration strategy in a self-healing distribution network with dg and flexible loads

  • A. Soroudi et al.

    Distribution network hosting capacity maximization using demand response

  • Cited by (60)

    • Reactive power potential of converter-connected renewables using convex power flow optimization

      2023, International Journal of Electrical Power and Energy Systems
    • Bi-level fuzzy stochastic-robust model for flexibility valorizing of renewable networked microgrids

      2022, Sustainable Energy, Grids and Networks
      Citation Excerpt :

      In other words, due to clean operation and low operating cost, they generally inject maximum output power as per climate changes to the network, leading to sharp power fluctuations [3]. Therefore, the intermittent nature of RESs in MG in terms of operating indices causes overvoltage, energy losses increase, MG lines overloading, etc. [4–6]. In addition, RESs together with other sources of uncertainty bring inconsistency between the results of day-ahead (DA) forecasting and real-time (RT) dispatch [7,8], where these challenges are generally addressed under the theme of reducing network flexibility [9].

    View all citing articles on Scopus
    View full text