(648e) A Superstructure Optimization Approach for the Design of Corn-Based Ethanol Plants

Peschel, A. - Presenter, Carnegie Mellon University
Karuppiah, R. - Presenter, Carnegie Mellon University
Grossmann, I. E. - Presenter, Carnegie Mellon University
Zullo, L. - Presenter, Cargill Incorporated
Martín, M. - Presenter, University of Salamanca

In this work, we address the problem of synthesizing corn-based bio-ethanol plants through the use of mathematical optimization and heat integration techniques. In such plants, fuel ethanol is produced using corn-kernels as the feedstock. Fuel grade ethanol has to be 100 % pure before it can be blended with gasoline to be used in automobiles. However, conventional distilleries produce an azeotropic mixture of ethanol and water (95 % ethanol ? 5 % water), which has to be purified further for making fuel ethanol. The main challenge in the way of producing fuel ethanol commercially is that the process is very energy intensive and requires large amounts of steam and electricity for use in the rectifiers to get an azeotropic mixture of ethanol and water, and further requires the use of expensive molecular sieves to get 100 % pure ethanol. Furthermore, the waste stillage from the fermentation units, which is rich in volatile organic compounds, and the other wastewater streams in the plant have to be treated before disposal, which, in turn requires energy intensive units like centrifuges and dryers. Due to the high costs involved in this process, producing fuel ethanol commercially is only now starting to become feasible with the high price of gasoline. If the economics of the process have to be improved, we would need to use process synthesis tools similar to those being used in the petrochemical industry to design optimally operating plants, rather than use ad-hoc or empirical approaches to build such systems. We look at the problem in a way to reduce the operating costs of the plant and minimize the energy usage and maximize the yields of the plant.

In order to design such a bio-ethanol plant, we propose a superstructure optimization approach where we first construct a fixed configuration flowsheet embedding the various process units involved in ethanol production. These units are interconnected to each other through network feed flows and other utility streams. Our objective is to optimize the structure, determining the connections in the network and the flow in each stream in the network, such that we minimize the energy requirement of the overall plant while trying to maximize the yields. For this purpose, heat integration and water re-use possibilities are included. We also try to exploit different options for feedstock and processing technologies. The optimization of the system is formulated as a Mixed Integer Non-linear Programming problem, where the model that is optimized involves mass and energy balances for all the units in the system. We optimize an example network for such a corn-ethanol plant and present results for the optimal configuration where the energy usage in the system is minimized. The results of the optimization are used in determining the hot and cold utilities required in the plant and heat integration is carried out. It is found that we are able to reduce the consumption of steam drastically by upto 70 % by using multi-effect columns for the distillation columns and performing heat integration. This in turn brings down the costs involved in producing a gallon of fuel grade ethanol.

Finally, it is worth mentioning that the production of such bio-ethanol can be a part of a larger installation known as a ?bio-refinery? where more than one biomass derived fuels and commodities are produced in a single facility. In such bio-refineries, plant biomass, which is the forseeable sustainable source of organic fuels and chemicals, would be used as the feed in the refinery and is converted to different commodities using biotechnology. The production of different co-products along with ethanol in a bio-refinery would improve the process economics and profitability of the plant.