There is a growing desire to optimize TEA models and understand the implications for real systems. Modern process modeling tools include optimization functions or can couple with stand-alone optimization software like IBM ILOG CPLEX Optimizer, GurobiTM, and GAMS among others. These tools allow researchers to systematically identify optimal operating parameters that meet certain constraints.

TEA models include parameters bounded by system constraints. For example, biorefineries include both technical (reactor temperature) and economic (minimum feedstock cost) constraints that need to be considered within the model. Within the bounds of the allowable parameters there are usually one or more function maxima or minima. In this regard, TEA systems are somewhat simpler than other mathematical models — the function space is well defined. The major challenges for optimization of TEA models are large, complex models with hundreds of parameters, and models that express extremely nonlinear behavior. Techniques that address both of these challenges are the subject of much research.

Researchers employ model surrogates or reduced order models (ROMs) to optimize large models that are either too complex or computationally expensive to evaluate. ROMs can significantly reduce the time required to optimize high-fidelity models at the risk of over-simplifying the problem. Therefore, several approaches have been proposed for the identification of ROM parameters and the evaluation of ROM accuracy.

The benefits of process optimization go beyond identifying optimal values. They also identify tradeoffs between differing objectives. These tradeoffs can be illustrated by a Pareto curve. Pareto curves describe the incremental changes of a given objective value due to improving a second objective. For example, biorefineries commonly face a tradeoff between lowering process costs from the use of fossil fuels and increasing their overall environmental footprint.

These emerging trends suggest a bright future for techno-economic analysis study and its impact on the advancement of biorefineries. The study of demonstration-plant data, combination of TEA and LCA, evaluation of risk and uncertainty, and optimization of system models are fertile grounds for future research and development.