(508d) Analysis of Parametric Uncertainty in An Advanced Coal Conversion Process Using the Probabilistic Collocation Method

Computer modeling plays an increasingly integral role the development of emerging technologies. Understanding the influence of parametric uncertainty is a crucial step before utilizing model predictions. Integrated gasification combined cycle (IGCC) systems represent a promising concept of power generation with integrated CO₂ capture capability, however, there remains significant uncertainty regarding the technical performance and costs. This uncertainty needs to be explicitly analyzed during the development of advanced technologies like IGCC, as it forms the basis for important decisions regarding design trade-offs, capital investments, and energy policy.

In this work, the Probabilistic Collocation Method (PCM) was used to characterize the performance of an IGCC model. PCM is a direct method of treating parametric uncertainty, using polynomial chaos expansions to represent the probability distribution functions [1]. This allows uncertainties from multiple parameters to be propagated through the model simultaneously. The PCM results are approximate probability distribution functions of the model outputs. This method can deal with large numbers of uncertain parameters within complex or even black-box models, such as AspenPlus flowsheets. Compared to Monte Carlo, PCM is more efficient, using orders of magnitude fewer model evaluations, and more informative, since PCM estimates the contribution to model output uncertainty from each uncertain parameter, while Monte Carlo only estimates the aggregate effect. This efficient treatment of uncertainty analysis also enables broader methods which incorporate model uncertainty, such as Model Based Experimental Designs.

A comprehensive process model of an IGCC system with CO₂ capture was developed using the AspenPlus simulator. The model is based on Case 2 of the NETL study on Fossil Fuel Power Plants [2]. It features a GE radiant cooler gasifier with oxygen supplied by an air separation plant. A two-stage water gas shift reactor system converts the CO-rich stream to a H₂-rich stream. Syngas is cooled in a steam generator to recover heat for power generation. Cool syngas then passes through a two-stage Selexol process to remove sulfur and CO₂. CO₂ is further compressed for geological sequestration. The process is designed to capture 95% of CO₂ in the syngas. The clean syngas, containing primarily H₂, is then combusted in a gas turbine combined cycle system. Key model outputs were: net power output, net plant efficiency, and relative CO₂ emissions.

Twenty-six IGCC model parameters, including: feedstock properties, design factors and operating conditions, were selected for analysis. Uncertainties were quantified and represented with probability distribution functions following a literature review and consultation with technical experts. PCM analysis on these parametric uncertainties on the three key outputs found considerable uncertainty in the net power output driven primarily by uncertainties in feedstock composition and gas turbine efficiency, while net plant efficiency and relative CO₂ emissions were less affected by input uncertainties.

Reference:

1. M. A. Tatang, W. Pan, R. G. Prinn, and G. J. McRae. An Efficient Method for Parametric Uncertainty Analysis of Numerical Geophysical Models. Journal of Geophysical Research, 102 (1997), pp. 21916- 21924.
2. Cost and Performance Baseline for Fossil Energy Plant, Vol. 1, DOE/NETL ? 2007/1281, (2007).