SC C6: Distribution Systems and Dispersed Generation Preferential Subject 2. Demand Management (DM) and Demand Response (DR) Role, willingness and approach of Distribution Network Operator in implementing DM&R projects A methodology based on Demand Side Management approaches for an optimal planning of MV-LV grid expansion in highly congested urban areas A. Capozza 1 , G. Morana 2 , F. Pilo 3 1 CESI SpA, Via Rubattino 54, 20134 Milano – Italy – email: capozza@cesi.it 2 Dipartimento di Ingegneria Elettrica – Università di Palermo – Italy – email: morana@diepa.unipa.it 3 Dipartimento di Ingegneria Elettrica ed Elettronica – Università di Cagliari – Italy – email: pilo@diee.unica.it 1 Foreword A renewed and growing interest is involving nowadays the Demand Side Management (DSM) techniques, since they are regarded as suitable tools in Distribution Networks planning. The following drivers trigger off this revitalisation: • The electric demand in urban contexts is growing beyond the expectations, owing to correspondingly increased needs of comfort (air-conditioning, indoor air quality, housekeeping, communication, etc.), which can be fulfilled through the electric end-uses with high quality, compliance with environmental regulations and competitive economic benefits. • This sudden higher pressure of the demand on the local Electric System is bound to force the Distributors to make huge short-term and unforeseen investments for expanding their urban MV-LV grids • The use of DSM-based tools and methodologies can slow down the growth rate of the yearly peak load and energy demand. Then, they can supply the planner with cheaper (and likewise efficient) alternatives with respect to a plain grid expansion. On these circumstances, the need is felt for the development of a methodology, which supplies the planners of urban grids with tools for a rational choice among several possible alternative mixes of investments and DSM measures, in order to fulfil the increased demand in an optimal way. A research project on this subject – ECORET – was set up by CESI under support of the Italian Fund for the Research for the Electricity System. The overall project outlined a praxis focused on three main aspects: 1. A set of techniques to evaluate the evolution of the load in highly crowded contexts such as the urban areas 2. A review on the cost/benefit effectiveness of typical DSM actions of either “load shedding” or “load levelling” kinds (such as cool storage, tri-generation for cool production, energy efficiency measures, household load automatic control) 3. A method for evaluating each time the most suitable mix of investments on grid expansion and of DSM measures, to face an unforeseen increase of the demand The present paper describes the adopted architectures and techniques, as well as the obtained results, relevant to the methodological aspect (point 3). The considered approach is made up of the two below described steps: preliminary screening and analysis on the real grid. 2 Preliminary screening and ranking of the DSM measures This former step consists of a propaedeutical screening and is carried on considering model generated grids. A reference scenario for the load growth is defined over a time horizon of about five years, with reference to a given distribution area in conditions of non-controlled demand (i.e. without any exploitation of DSM measures). An economic ranking is defined among a suitable number (even a large one, tentatively 15-20) of alternative mixes of possible DSM measures and investments in grid developments. A corresponding number of scenarios for the load growth is evaluated as well, with reference to the same area and to each one of the above considered DSM/investment mixes (DSM driven scenarios).

A minimal subset of DSM/investments mixes (tentatively, 3-5) is selected as particularly effective and efficient, to be considered in the further Step 2 for analysis on the real grid. The efficiency index is given by a comparison between benefits and costs, according to the following criteria. • The benefits are those arising from lower or deferred investments for grid adjustment, as a consequence of the slower increase of the peak load brought about by the DSM measures against the reference scenario. These avoided costs are evaluated by the use of the computer programme DECAPLAN. This software evaluates the grid development by operating on model generated grids; it was developed by the Department of Electrical Engineering of the University of Palermo (I). • The costs are referred to the launch of the DSM measures and to possible non-deferrable investments 3 Analysis of DSM effects on the real grid The present step 2 is carried on considering the real grid and the minimal subset of DSM measures resulting from the previous step. It consists of three sequential phases: 1. Adjustment, reorganisation and optimisation of the real urban grid against the reference scenario of load growth (in absence of DSM measures) and evaluation of the involved costs 2. Adjustment, reorganisation and optimisation of the real urban grid against the subset of the DSM driven scenarios of load growth resulting from the previous step and evaluation of the involved costs. In this phase, the chance of exploiting the grid more intensively and with tighter margins against the standard tolerances (the so-called relaxation of technical constraints) is considered, in combination with undertaking DSM policies; to this purpose, the correspondent higher charges deriving from a worse quality of power supply must be kept into account. 3. Evaluations of the efficiency of the considered DSM measures by comparison of benefits and costs: • Avoided costs, resulting from the differential between costs in reference scenario (see point 1.) and in DSM-driven scenarios (see point 2.) • Costs connected to the undertaken DSM measures • Charges following from relaxation of technical constraints (e.g. penalties to be paid to regulatory Authorities, discounts granted to the clients against a decrease of quality of power supply, advanced investments for a faster ageing of the machinery, etc) The evaluation on the real grid relies on the use of the computer programme SPREAD, jointly developed by the Department of Electrical and Electronic Engineering of the University of Cagliari (I) and CESI. This software is aimed at: • MV grid planning through its economic optimisation, in compliance with technical constraints and service continuity requirements • Definition of the technically and economically optimal chronology of grid developments to attain its final configuration. 4 Case study The effectiveness of the above methodology is verified on a case study. The study is relevant to a MV grid characterised by: • A served surface of about 9 sqkm, characterised by a composition of users and by a load density typical of highly populated urban areas • Three HV/MV substations, which feed a complex network of about 100 MV feeders; each MV line is radially operated; disconnectors are installed along the feeders to allow backfeeding configuration • A number of about 1500 nodes, including MV users and MV/LV substations; these substations on their turn feed a grid of about 700 LV feeders • A total peak load during the reference year of about 260 MVA; 144 MVA of this peak load are to be ascribed to MV/LV substations. • Growth rates of 2 %/year for the overall peak load and of 1 %/year for the LV peak load were assumed in absence of DSM policies (i.e. in the above reference conditions) The preliminary screening of the first step reviews about 10 typical DSM measures. An intensive use of DECAPLAN programme on a model generated grid leads to the selection of two-three of these measures as the most cost effective.

The evaluations on the real grid made in the second step is performed through the use of the SPREAD software considering the previously selected DSM measures. These studies allow for a detailed evaluation of the costs and benefits involved by the selected measures. These studies also give information on the technical/economic effectiveness of possible actions based on relaxation of some technical constraints.