A framework for quantifying the value of information to mitigate risk in the optimal design of distributed energy systems under uncertainty

Abstract

Distributed energy systems (DESs) are regarded as promising systems for integrating renewable energy sources. However, uncertainties arising from renewable energy and loads introduce significant complexity to DES design and may even result in reliability and economic risks when the design of DESs relies on limited information. Gathering more information can reduce uncertainty, thereby improving the robustness of the DES scheme. However, obtaining information comes at a cost, and too much information can result in redundant work and unnecessary computing burden. Conversely, discarding or ignoring information may pose risks to reliability and the economy. Therefore, this study presents a framework for quantifying the value of uncertainty information, which can help to understand how information affects risk and identify key information that facilitates DES risk aversion. Two information value indices, namely the expected values of information for reliability (EVPIr) and economy (EVPIe), are developed to measure the risk reduction of reliability and economy when more information is added to the design of DESs. Furthermore, a two-layer information value quantification model based on mixed integer linear programming is built to optimize the design of DESs based on uncertain information and quantify the value of information based on a relatively complete information set. The proposed information value quantification method is tested on a real DES under three types of uncertain design boundary scenarios. The results show that the values of EVPIr and EVPIe decrease with increasing information of uncertain design boundary scenarios, indicating that more information reduces risks. An unexpected discovery is that the probability information of the scenario set is not critical for DESs. The deviations of EVPIe are within ±2%. The proposed approach offers a quantitative means to evaluate and filter key information for planning scenarios, which can facilitate the generation of streamlined planning scenarios without compromising reliability and economy.