Impacts of Ramping Inflexibility of Conventional Generators on Strategic Operation of Energy Storage Facilities

Abstract

Abstract-This paper proposes an approach to assist a pricemaker merchant energy storage facility in making its optimal operation decisions. The facility operates in a pool-based electricity market, where the ramping capability of other resources is limited. Also, wind power resources exist in the system. The merchant facility seeks to maximize its profit through strategic inter-temporal arbitrage decisions, when taking advantage of those ramp limitations. The market operator, on the other hand, aims at maximizing the social welfare under wind power generation uncertainty. Thus, a stochastic bi-level optimization model is proposed, taking into account the interactions between the storage facility and the market operator, and the existing market opportunities for the storage facility. The proposed bilevel model is then transformed into a Mathematical Program with Equilibrium Constraints (MPEC) that can be recast as a Mixed-integer Linear Programming (MILP) problem. Different case studies are presented and discussed using a six-bus illustrative example and the IEEE one-area reliability test system to evaluate the performance of the proposed approach. Capacity of conventional generator g, (MW) P ini gInitial value of generated power of conventional generator g, (MW) RU

Incorporating Non-Convex Operating Characteristics Into Bi-Level Optimization Electricity Market Models

“…Bi-level optimization problems are usually solved after converting them to single-level Mathematical Programs with Equilibrium Constraints (MPEC), through the replacement of the lower level problem by its equivalent Karush-Kuhn-Tucker (KKT) optimality conditions. This popularity is justified by the large number of papers employing this methodology, including papers [2]- [25]. This list includes papers focusing on market models including only traditional generation participants [2]- [16] as well as models investigating the impact of demand side flexibility [17]- [20] and energy storage [21]- [25] in electricity markets.…”

“…This popularity is justified by the large number of papers employing this methodology, including papers [2]- [25]. This list includes papers focusing on market models including only traditional generation participants [2]- [16] as well as models investigating the impact of demand side flexibility [17]- [20] and energy storage [21]- [25] in electricity markets. Furthermore, some of these papers focus on the optimization of the strategic bidding decisions of a single market player [5], [8], [9], [12], [15], [16], [18], [19], [21]- [25], while others have gone further, aiming at identifying the market equilibria stemming from the interactions of multiple players [2]- [4], [6], [7], [10], [11], [13], [14], [17], [20].…”

“…This list includes papers focusing on market models including only traditional generation participants [2]- [16] as well as models investigating the impact of demand side flexibility [17]- [20] and energy storage [21]- [25] in electricity markets. Furthermore, some of these papers focus on the optimization of the strategic bidding decisions of a single market player [5], [8], [9], [12], [15], [16], [18], [19], [21]- [25], while others have gone further, aiming at identifying the market equilibria stemming from the interactions of multiple players [2]- [4], [6], [7], [10], [11], [13], [14], [17], [20]. Finally, some of these papers have managed to incorporate realistic aspects such as network constraints [2]- [6], [9], [11], [12], [14]- [18], [20]- [25], and uncertainty faced by market players [5], [8]- [10], [12]- [16], [18], [19], [24], [25] in the developed market models.…”

“…This is that the lower level problem does not include any binary decision variables since the derivation of the equivalent KKT optimality conditions is only possible when this problem is continuous and convex [26]. As a result, the models developed in all previous relevant papers [2]- [25] consider only the variable costs, maximum output limits and ramp rates of generation units and they neglect physical non-convex cost components and constraints that are associated with binary commitment decisions, including no-load, start-up and shutdown costs, minimum stable generation limits, and minimumup / down time constraints.…”

“…Two recent pieces of work have proposed mathematical approaches to solve bi-level optimization problems with binary variables in the lower level problem, although they have applied these approaches to fundamentally different problems than the strategic bidding problem examined in papers [2]- [25] and this paper. The first one involves an iterative solution technique based on a column-and-constraint generation method [32], [33].…”