Uncertainty & sensitivity analysis of economic assessment of lactic acid production from crude glycerol - impact of price correlations

In this work, we investigate the impact of the expected price volatility and correlations on the overall economic assessment of lactic acid production from crude glycerol. In particular, the goals of this study are three-fold: (i) to understand how uncertainty in the inputs propagates to the model outputs; (ii) to understand the effect of the degree of pairwise correlation between input uncertainties on each other and on the outputs from the economic model (Net Present Value); and lastly, (iii) to estimate the first-order as well as independent sensitivity indices so as to identify which of the input uncertainties in the economic analysis affect the estimated NPV the most. To this end, we implemented algorithms in Matlab (R2015a) based on Monte Carlo sampling with permutation using Latin Hypercube Sampling with Iman Conover correlation control (Sin et al., 2009).The results have shown that the estimated NPVs are very sensitive to uncertainties in the inputs. In particular, not only the magnitude of the input uncertainty but also the degree and the sign of the correlation among input uncertainties matter a lot. All in all, this approach aims at providing information and powerful insights on the quality of the estimated NPV, a commonly used economic indicator, for assessing high risk and high gain biotech process investments. The results help to better assess economic feasibility under a broad range of uncertainties and ultimately minimize the risk of potential business failure commonly seen in past biorefinery developments.