Sanjoy Das

M. Nazif Faqiry, Ahmad Khaled Zarabie, Fatehullah Nassery et al.

In this paper, we study a day ahead double energy auction in a distribution system involving dispatchable generation units, renewable generation units supported by battery storage systems(BSSs), fixed loads, price responsive loads, and supply from the Whole Sale Market(WSM) at Locational Marginal Price(LMP). The auction is implemented within a Distribution System Operator (DSO) premises using Mixed Integer Linear Programming (MIP). The proposed auction is cleared at the Distribution LMP (DLMP) and is observed to be weakly budget balanced if no penalty is applied for DSO's deviation from originally committed supply from the WSM. Furthermore, the dynamics of LMP and DLMP, and their effect on distribution market participants scheduled quantities as well as the WSM supply to the distribution system is investigated.

Mohammad Nazif Faqiry, Sanjoy Das

A novel distributed energy allocation mechanism for Distribution System Operator (DSO) market through a bi-level iterative auction is proposed. With the locational marginal price at the substation node known, the DSO runs an upper level auction with aggregators as intermediate agents competing for energy. This DSO level auction takes into account physical grid constraints such as line flows, transformer capacities and node voltage limits. This auction mechanism is a straightforward implementation of projected gradient descent on the social welfare (SW) of all home level agents. Aggregators, which serve home level agents - both buyers and sellers, implement lower level auctions in parallel, through proportional allocation and without asking for utility functions and generation capacities that are considered private information. The overall bi-level auction is shown to be efficient and weakly budget balanced.

A. Khaled Zarabie, Sanjoy Das, Hongyu Wu

While non-parametric models, such as neural networks, are sufficient in the load forecasting, separate estimates of fixed and shiftable loads are beneficial to a wide range of applications such as distribution system operational planning, load scheduling, energy trading, and utility demand response programs. A semi-parametric estimation model is usually required, where cost sensitivities of demands must be known. Existing research work consistently uses somewhat arbitrary parameters that seem to work best. In this paper, we propose a generic class of data-driven semiparametric models derived from consumption data of residential consumers. A two-stage machine learning approach is developed. In the first stage, disaggregation of the load into fixed and shiftable components is accomplished by means of a hybrid algorithm consisting of non-negative matrix factorization (NMF) and Gaussian mixture models (GMM), with the latter trained by an expectation-maximization (EM) algorithm. The fixed and shiftable loads are subject to analytic treatment with economic considerations. In the second stage, the model parameters are estimated using an L2-norm, epsilon-insensitive regression approach. Actual energy usage data of two residential customers show the validity of the proposed method.