A Decision Theory Approach to Cogeneration Planning in the Presence of Uncertainties

This paper presents a general approach to cogeneration planning in a multi-year time interval of analysis, taking into account various types of uncertainties. The approach is structured according to different time frames. Starting from the electrical and thermal load patterns for a reference period, consistent groups of hours are created in the short-term time frame. The relevant energy and electricity price random variables (RVs) are represented through average values and covariance matrices in order to account for their possible correlations. Thus, the RVs are extracted from multivariate Normal probability distributions, which allows for Monte Carlo simulations of the cogeneration plant operation within each group of hours. The evolution of fuel prices and financial parameters is set in the medium-term time frame. To account for the presence of uncertainties, decision theory-based criteria are then applied in the long-term time frame to select the alternative that presents the best performance over a set of scenarios. A case study application, based on the data referred to a hospital site, shows the viability of the proposed approach. The planning alternatives include technologies such as microturbines and internal combustion engines of different sizes. The scenarios are characterized by specific evolution of the electrical and thermal loads and of the electricity and gas prices. The mean value and the 5% non-exceeding probability of the net present costs are used as decision variables.