Towards Robust Scenarios of Spatio-Temporal Renewable Energy Planning: A GIS-RO Approach

Solar-based energy is an intermittent power resource whose potential pattern varies in space and time. Planning the penetration of such resource into a regional power network is a strategic problem that requires both to locate and bound candidate parcels subject to multiple geographical restrictions and to determine the subset of these and their size so that the solar energy production is maximized and the associated costs minimized. The problem is also permeated with uncertainty present in the estimated forecast energy demand, resource potential and technical costs. This paper presents a novel combination of Geographic Information Systems (GIS) and Robust Optimization (RO) to develop strategic planning scenarios of a collection of parcels that accounts for their spatio-temporal characteristics, and specifically their hourly radiation patterns that are location dependent, to best fit the network temporal demand and minimize technical costs. The problem is formulated as a GIS spatial placement problem and a RO fractional knapsack problem to plan the effective power penetration and geographical suitability of new PV facilities. The combination GIS-RO generates an excellent decision support system that allows for the computation of optimized parcel scenarios (locations, sizes and power). The qualitative and quantitative effectiveness of the approach is demonstrated on real data on the French Guiana region. Results show that the proposed approach provides reliable fine grained planning that also accounts for the risk adversity of the decision maker towards forecast demand and solar potentials.