Contents lists available at ScienceDirect Electric Power Systems Research journal homepage: www.elsevier.com/locate/epsr An integrated day-ahead market clearing model: Incorporating paradoxically rejected/accepted orders and a case study Sinan Yörükoğlu a, ⁎ , Zeynep Müge Avşar b , Bora Kat c a AF-Mercados EMI, Mustafa Kemal Mah. Dumlupınar Blv. No: 280, D Blok, No:3, 06510 Ankara, Turkey b Department of Industrial Engineering, Middle East Technical University, Ŭniversiteler Mah. Dumlupınar Blv. No:1, 06800 Ankara, Turkey c The Scientific and Technological Research Council of Turkey, Tunus Cd. No:80, 06100 Ankara, Turkey ARTICLE INFO Keywords: OR in energy Electricity day-ahead market Combinatorial auctions Social welfare Non-linear programming ABSTRACT In the electricity day-ahead markets (DAMs), market participants place their orders in the form of desired/ accepted price levels for the submitted quantities. The market operator evaluates these orders and announces the clearing quantities and market clearing prices (MCPs) within an hour. In this paper, a mathematical model is proposed for the exact solution of clearing electricity DAM with a focus on the current Turkish system. The model is a non-linear programming (NLP) problem that maximizes total social welfare and takes into account all types of orders that are submitted in the Turkish DAM. In order to ensure a feasible solution, the concept of paradoxically accepted/rejected orders is introduced. Two versions of the mathematical model are considered, allowing for one type of paradoxical processing in each version. For the computational experiments, a sample data set of 25 days, representing the conditions in the Turkish DAM, is generated and published on an open library. The model is solved to optimality within the practical time limitation of the market operator in all cases. 1. Introduction Electricity can be traded in many different ways. In general, it can be accomplished either by long-term power purchase agreements (PPAs) or the spot market based on short-term auctions and energy exchange mechanisms. In any case, as noted in [1], the supply and demand must be always in balance as electricity is an intangible com- modity that is difficult to store. As compared to the European markets, in Turkey greater amount of electricity is traded under long-term PPAs while the spot market, which is called by law the balancing and set- tlement market (BSM), has also a significant volume. The main platform to trade electricity in the BSM is the day-ahead market (DAM). It is reported in [2] that the share of DAM in total electricity trade carried out in Turkey was around 42% and 43% in 2016 and 2017, respec- tively. This paper reviews the market clearing approaches in European markets with a specific focus on Turkey. European spot markets are organized as power exchange (PX) markets, which take place in the form of double auctions (double blind auctions), meaning that there are several bidders on both buyer and seller sides for the same product or commodity [3]. The Central Western European (CWE) region couples the PX markets of five countries (Belgium, France, Germany, Lux- embourg and the Netherlands) and clears those markets by making use of a Branch-and-Bound based algorithm called COSMOS [4]. Similarly, the Scandinavian and Baltic countries (Finland, Denmark, Sweden, Norway, Estonia and Lithuania) used to be coupled by Nord Pool Spot coupling system called SESAM [5]. The EPEXSpot [6] integrated the PX markets of the United Kingdom, Austria and Switzerland with the CWE countries. Finally, seven European PX markets have been merged by the system called “Pan-European Hybrid Electricity Market Integration Algorithm (EUPHEMIA)” [7], which covers 23 individual countries. 1 • Before EUPHEMIA, for instance, the clearing process in the Iberian peninsula (Spain and Portugal) consisted of several stages im- plemented by OMEL [8]. These stages are the calculation of a pro- visional clearing price involving only simple bids (hourly bids having only a price and amount of energy), and step by step, iterative inclusion of complex bids, 2 such as Load Gradient, In- divisibility, Minimum Income and Scheduled Shutdown bids. • The COSMOS algorithm [4] used in CWE region includes single https://doi.org/10.1016/j.epsr.2018.07.007 Received 21 March 2018; Received in revised form 13 June 2018; Accepted 6 July 2018 ⁎ Corresponding author. E-mail addresses: syyorukoglu@gmail.com (S. Yörükoğlu), mavsar@metu.edu.tr (Z.M. Avşar), borakat@gmail.com (B. Kat). 1 Austria, Belgium, Czech Republic, Denmark, Estonia, Finland, France, Germany, Hungary, Italy, Latvia, Lithuania, Luxembourg, the Netherlands, Norway, Poland, Portugal, Romania, Slovakia, Slovenia, Spain, Sweden and UK. 2 Complex bid is a special type of simple bids in the sales direction that contain various “complex” conditions, creating additional restrictions on hourly bids. Electric Power Systems Research 163 (2018) 513–522 Available online 02 August 2018 0378-7796/ © 2018 Elsevier B.V. All rights reserved. T