The Power & Energy Society (PES) General Meeting will be held in Boston from July 17th-21st 2016. Energy Institute UCD will be well represented at the conference with the following presentations or posters being given:

Validating Unit Commitment Models: A Case for Benchmark Test Systems -Alexander C. Melhorn, Mingsong Li, Paula Carroll, and Damian Flynn

Due to increasing penetration of non-traditional power system resources; e.g. renewable generation, electric vehicles, demand response, etc. and computational power there has been an increased interest in research on unit commitment. It therefore may be important to take another look at how unit commitment models and algorithms are validated especially as improvements in solutions and algorithmic performance are desired to combat the added complexity of additional constraints. This paper explores an overview of the current state of unit commitment models and algorithms, and finds improvements for both comparing and validating models with benchmark test systems. Examples are provided discussing the importance for a standard benchmark test system(s) and why it is needed to compare and validate the real world performance of unit commitment models.

“Hierarchical Demand Response for Peak Minimization Using Dantzig–Wolfe Decomposition”, Paul Mc Namara, Seán McLoone

Demand  response  (DR)  algorithms  manipulate  the energy consumption schedules of controllable loads so as to satisfy  grid  objectives.  Implementation  of  DR  algorithms  using  a centralized agent can be problematic for scalability reasons, and there are issues related to the privacy of data and robustness to communication  failures.  Thus,  it  is  desirable  to  use  a  scalable decentralized  algorithm  for  the  implementation  of  DR.  In  this paper, a hierarchical DR scheme is proposed for peak minimization based on Dantzig–Wolfe decomposition (DWD). In addition, a time weighted maximization option is included in the cost function, which improves the quality of service for devices seeking to receive their desired energy sooner rather than later. This  paper also demonstrates how the DWD algorithm can be implemented more  efficiently  through  the  calculation  of  the  upper  and  lower cost bounds after  each  DWD  iteration.

“Design of MPC-based Controller for a Generalized Energy Storage System Model” – Álvaro Ortega, Paul Mc Namara, Federico Milano

This paper presents a control strategy based on Model Predictive Control for Energy Storage Systems. The mathematical formulation of this controller is outlined, and the procedure for applying this controller to a Generalized Energy Storage model is then documented. The dynamic performance of the control strategy presented is compared with that of a PI -based control technique. A comprehensive case study based on the New England 39-bus 10-machine test system with the
inclusion of Energy Storage Systems is presented and discussed.

 “Model Predictive Control based AGC for Multi-Terminal DC grids” – Paul Mc Namara, Álvaro Ortega, Federico Milano

With increasing DC grid connections between non-synchronous AC systems it is desirable that DC connections would take a role in frequency regulation for connected AC grids. A number of primary and secondary P and PI based controllers have been designed previously for this purpose. Here Model Predictive Control is proposed for including DC power controllers in the provision of Automatic Generation Control.

 “Information gap decision theory based congestion and voltage management in the presence of uncertain wind power”, Conor Murphy

The supply of electrical energy is being increasingly sourced from renewable generation. The variability and uncertainty of renewable generation, compared to a dispatchable plant, is a significant dissimilarity of concern to the traditionally reliable and robust power system. This change is driving the power system towards a more flexible entity that carries greater amounts of reserve. For congestion management purposes it is of benefit to know the probable and possible renewable generation dispatch, but to what extent will these variations effect the management of congestion on the system? Reactive power generation from wind generators and demand response flexibility are the decision variables here in a risk averse multi-period AC optimal power flow (OPF) seeking to manage congestion on distribution systems. Information Gap Decision Theory is used to address the variability and uncertainty of renewable generation. In addition, this work considers the natural benefits to the congestion on a system from the over estimation of wind forecast; providing an opportunistic schedule for both demand response nodes and reactive power provision from distributed generation.

 “Transmission Planning, Flexibility Measures and Renewables Integration for Ireland Power System”, Damian Flynn

The Irish power system has proposed challenging renewable energy (mainly wind generation) targets for 2020. Given that much of the wind generation will be connected at distribution level, network development at both distribution and transmission level is a key requirement for both the transmission system operators (TSO) and distribution system operators (DSO) on the system. This has involved the analysis of a wide range of technical options, including undergrounding, DC connections and series compensation options, supported by a co-ordinated public engagement programme. Incentivising system flexibility in the form of new ancillary services associated with system ramping, synchronous inertia and dynamic reactive power are also being pursued, and are being pushed towards implementation. The experience learned from the Irish system is highlighted, indicating the solutions which have ultimately been proposed, and remaining challenges for the future.


Residential Load Modeling of Price Bases Demand Response for Network Impact Studies, Killian McKenna

This paper presents a comprehensive low-voltage residential load model of price based demand response. High- resolution load models are developed by combing Monte Carlo Markov Chain bottom-up demand models, hot water demand models, discrete state space representation of thermal appliances and composite time-variant electrical load models. Price based demand response is then modeled through control algorithms for thermostatically controlled loads, optimal scheduling of wet appliances and price elasticity matrices for representing the inherent elastic response of the consumer. The developed model is used in a case study to examine the potential distribution network impacts of the introduction of dynamic pricing schemes. The effects of cold load pick-up, rebound peaks, decrease in electrical and demand diversity and impacts on loading and voltage are presented.