Search

This paper presents two mathematical formulations to represent uncertainties in self-scheduling models of a price-taker Compressed Air Energy Storage (CAES) facility. The proposed model is from the point of view of the plant owner participating in the energy, spinning, and idle reserve markets. The first described formulation is based on Robust Optimization (RO) and the second one is based on Affine Arithmetic (AA) techniques, which are both range arithmetic methodologies, and consider the thermodynamic characteristics of the CAES facility for a more realistic representation. The implementation of both methods are tested, validated and compared with each other and with Monte Carlo Simulations (MCS) using prices from the Ontario market. From the simulation results, it can be observed that both methods have some similarities, presenting lower computational burden compared with MCS, and demonstrate the advantage of applying the proposed models for CAES plant owners to hedge against price uncertainties. [ABSTRACT FROM AUTHOR]

Energy storage systems (ESSs) can improve energy management in distribution grids, especially with the increasing penetration of home energy management systems (HEMSs) that schedule household appliances and render them as smart loads. A large number of uncoordinated HEMSs can result in significant changes to the aggregated load profile of the distribution system. This paper proposes a framework and mathematical model for integrating ESS in the distribution grid to minimize the operation cost of the local distribution company (LDC) and alleviate the impact of uncoordinated HEMS operation on the distribution grid. A novel neural network (NN) based state of health (SoH) estimator for a lithium-ion (Li-ion) battery based ESS is proposed, which is incorporated within the LDC's planning problem. The results show that the proposed estimation model is an accurate estimation of the SoH of the ESS. The LDC's planning decisions are also compared, considering SoH of the ESS vis-á-vis linear degradation and no-degradation models. [ABSTRACT FROM AUTHOR]

Oil refining is an energy-intensive industry, that is often equipped with different types of distributed generation (DG) resources. Optimal energy management of the refinery's load and its DG resources, under dynamic electricity pricing, improves electricity consumption behavior of the facility and hence reduces its energy costs. An energy management system (EMS) model is proposed in this paper for minimizing electricity consumption costs of an oil refinery facility considering on-site cogeneration capability. The energy management is based on power demand modeling of the oil refinery process. A joint electrical–thermal model is used for the cogeneration units to account for the electricity and steam production. The developed EMS model is also used as part of a demand response strategy to illustrate the impact of EMS decisions on distribution system operations. [ABSTRACT FROM AUTHOR]

With the increasing importance of battery energy storage systems (BESS) in microgrids, accurate modeling plays a key role in understanding their behavior. This paper investigates and compares the performance of BESS models with different depths of detail. Specifically, several models are examined: an average model represented by voltage sources; an ideal dc source behind a voltage source converter; a back-to-back buck/boost and bidirectional three-phase converter, with all models sharing the same control system and parameters; and two additional proposed models where the switches are replaced by dependent sources to help analyze the differences observed in the performance of the models. All these models are developed in PSCAD and their performances are simulated and compared considering various issues such as voltage and frequency stability and total harmonic distortion in a benchmark test microgrid. It is shown through simulation results and eigenvalue studies that the proposed models can exhibit a different performance, especially when the system is heavily loaded, highlighting the need for more accurate modeling under certain microgrid conditions. [ABSTRACT FROM AUTHOR]

This paper presents a novel probabilistic distribution load flow (PDLF) algorithm to study the effect of connecting a wind turbine (WT) to a distribution system. This probabilistic approach is used to capture the stochastic behavior of the generation of WTs. Three different models of WTs are developed to be embedded in the PDLF. Consequently, a probabilistic approach to evaluate the impact of wind penetration into distribution systems is developed. Futhermore, the effect of WT penetration on feeder losses, voltage profile and substation powers is presented. [ABSTRACT FROM PUBLISHER]