Matteo Saviozzi

Distributed Energy Resources Management in a Low-Voltage Test Facility

The electric energy demand will increase in the future, and the will to exploit larger amounts of generation from renewable resources requires the development of new strategies to manage a more complex electrical system. Different techniques allow the smart management of distribution networks such as load shifting, peak shaving, and short-term optimization. This work aims to test, in a real low-voltage (LV) active network (LV test facility of Strathclyde University of Glasgow), a Microgrid Smart Energy Management System, which adopts a two-stage strategy. The two levels of the proposed energy control system are composed of: 1) midterm controller that, according to weather, load, and generation forecasts, computes the profile of the controllable resources (generation, load, and storage), the dispatch problem is then solved through an optimization process; and through 2) short-term controller, which controls the power absorption of the active network. This procedure is hierarchically designed to dispatch the resources/loads, according to priority signals with the objective to contain the energy consumption below predetermined thresholds. The scalability and effectiveness of the architecture, which is validated in a real test bed, demonstrates the feasibility of implementing such a type of controller directly connected to the LV breakers, delivering a part of a real smart grid.

Performance evaluation of algorithms for the State of Charge estimation of storage devices in microgrid operation

This paper analyzes different Kalman filtering algorithms for the real-time State of Charge (SoC) estimation of Battery Energy Storage System (BESS). Accurate SoC estimation is a key issue for microgrid real-time operation involving optimal model-based control. A BESS composed of Li-ion battery equipped with a Battery Management System (BMS) is characterized by fitting the parameters of a dynamic model, validated through experimental tests. Particular attention is devoted to the identification and representation of model nonlinearities in order to design robust Kalman filtering SoC estimation methods. Performance evaluation of the proposed algorithms are carried out by statistical simulations and experimental real-time tests. The analysis also takes in consideration the computational performances of the different methods in order to match the requirements of real-time control routines.

An architecture for implementing state estimation application in Distribution Management System (DMS)

This work aims to propose a state estimation procedure for electric distribution networks and to implement a simulation architecture in order to test it under several scenarios. Measures, obtained from the simulated field through a communication layer, and pseudo-measures, defined according to load and generation models, are combined in different shares to observe the algorithm performances and the effects on the estimation quality. The work also investigates the necessity of a correct load modeling of Medium Voltage/Low Voltage (MV/LV) substations where distributed generation significantly contributes in altering the net power injection at the LV side. The implemented State Estimation (SE) procedure will be installed in a real MV network located in the North of Italy (Sanremo) within a research project.

Smart management of demand in naval application: Prospects and technologies for distributed control of loads

The paper aims to introduce an interesting concept of Power Management System, which adopts a full distributed architecture for achieving Smart Management of loads. An effective way to prepare and manage electric network with the increased demand, is to apply all the different techniques available in smart grids, such as load balancing, load shifting and peak shaving through the intelligent management of the load (ILM). This approach can be also used on vessel in order to perform different specific objectives in accordance to the PMS (Power Management System) or as a back-up solution of this component.

Mixed-integer algorithm for optimal dispatch of integrated PV-storage systems

The exploitation of combined photovoltaic (PV) plants and storage systems is nowadays assuming growing importance, due to the technical, environmental, and economical benefits which can derive from an optimal integration. In this paper, a mixed-integer algorithm for the optimal dispatch of a storage system, based on the day-ahead PV forecasting is developed. The optimization objective is the maximization of the total production of the integrated system, according to a requested active power profile, which can be defined by the operator. The study case of an existing distribution management system, which operates on the low-voltage microgrid at University of Genova is analyzed. The procedure is validated by field results with particular attention to the storage round-trip efficiency.

Real time control of a low voltage microgrid through a Distribution Management System

The technological, economical and social changes of the last years had an impressive impact in the electrical energy world. The improvement in information and communication technologies, the liberalization of the energy market, the appearance of active customers and the growing spread of renewables have required to systems operators to recognize and manage new scenarios. Nowadays, it is globally accepted that distribution systems are the core of this change, which will lead to the realization of the smartgrid concept. As the key field of the transition is represented by the distribution system, the main role is played by Distribution System Operators (DSOs). They are asked to modernize their medium and low voltage networks and to adopt advanced control and management systems. Modern Distribution Management Systems (DMSs) answer this request, by providing a valuable and effective support to system operators, thanks to advanced functionalities, embedded on board, which support the exploitation of unpredictable resources (as renewable generation), controllable devices (distributed generators and storage systems) or active customers. In this context this paper aims to describe the implementation of a DMS for the real time control of a low voltage microgrid, focusing on the communication architecture and on the routine of the embedded advanced functionalities.

Mixed-Integer Linear optimization algorithm for Volt/Var Control on a distribution grid with renewable penetration

Nowadays, the diffusion of renewable energy sources and electric vehicles are turning the traditional passive distribution networks into active grids, requiring Distribution Management System (DMS) to exploit Information and Communication Technology (ICT) improvements to get a smarter control through wider communication and local monitoring. In this paper, a Mixed-Integer Linear Programming (MILP) algorithm is developed to perform Volt/Var Optimization (VVO), providing the optimal setting for all the available controllable devices which minimizes power losses and violations of network constraints. Necessary linearization affects acceptably the accuracy of the computed outcomes and the resulting fast computational time makes it suitable for online application within a real DMS.

Probabilistic assessment of Net Transfer Capacity considering forecast uncertainties

In transmission system planning, researchers propose methods to assess the effect of uncertainties of power system operating condition due to forecasting errors of intermittent generation and loads. In particular probabilistic power flow methods are used to calculate the probability distributions of the voltages and the branch currents, starting from the distributions of power injections/absorptions. These uncertainties play a key role in the operational planning of power systems, as certain configurations of load and intermittent generation can cause security problems. This paper aims to propose a probabilistic methodology to assess Net Transfer Capacity (NTC) among network areas, which quantifies forecast error uncertainties by applying the Point Estimate Method (PEM) combined with Third Order Polynomial Normal (TPN) Transformation. This approach is compared with a conventional NTC assessment technique and has been tested on an IEEE test system.