Aspen plus simulation of starch production by corn wet milling process

Abstract

The increasing public awareness on environmental issues and tighthening regulations are driving the chemical industry towards "Sustainable Business Models" that prioritize reducing fossil resource consumption and preventing waste. However, the necessary transition has been slower than required, making the efficient use of relatively more environmentally friendly conventional biotechnological processes crucial. Accordingly, as an established and profitable biotechnological process, we have identified corn wet milling (CWM) as an ideal platform for implementing modern approaches, such as process integration, to enhance the efficiency and sustainability of chemical industry. Nevertheless, to the best of our knowledge, a comprehensive process simulation of CWM is not available in open literature. This gap in the literature motivated us to perform the present study. CWM is an essential industrial process used for separating corn into its valuable components, including starch, gluten, fiber, and germ. This process plays a critical role in producing a wide range of products, such as corn sweeteners, corn oil, and various food and industrial ingredients. To explore the integration of CWM with various systems or to assess the profitibility potential of various process optimization strategies, it is critical to simulate the process accurately. To address this gap, we simulated CWM in Aspen Plus, the leading computational simultion tool offering advanced thermodynamic and transport models, and comprehensive biocomponent data. Corn kernels, the primary raw material input for the process, have a complex and heterogeneous structure that makes it impossible to characterize the feed stream in Aspen Plus as a straightforward conventional solid. Since no detailed data is available in the Aspen database, we backcalculated the corn kernel composition through the end products in CWM. Using proximate and ultimate analyses data, we have defined starch and lignin as nonconventional solids; and we have considered the remaining components as a mixture of their respective structural monomers. Notably, we have employed the Van Krevelen methodology for functional group identification. Moreover, to reliably predict the thermodynamic and transport properties, we employed the NRTL model. We have simulated the process as accurately as possible, largely based on the know-how of Sigma Process Technologies, a Türkiye-based process design company with a worldwide leading position in CWM. To better represent the actual production process in the simulation environment, we paid extra attention to accurately define the particle size distributions. To reflect the particle size changes that occur during the release of starch and gluten granules embedded in the endosperm under hydraulic and mechanical forces, such as in hydrocyclones, we incorporated grinding units into the simulation, although they do not exist in the actual process. Another significant challenge in modeling was achieving effective phase separation driven by density differences. Due to the inability to specifically define components like germ and fiber within the simulation environment, we could not achieve a perfectly ideal separation performance. To overcome this limitation, we combined hydrocyclone and centrifuge blocks with a separator and treated the combination as an integrated system. Other than these, we have reflected the actual process as faithfully as possible. Eventually, we yielded the following product distribution (on a dry basis): ~51.7% starch, ~10.6% gluten, ~11.2% fiber, ~8.6% germ, and ~13.7% light steep water (LSW). Given that the results approximated to the actual values with a deviation as small as 30%at most, except for LSW, we have made substantial progress in constructing a reliable simulation of CWM. Finally, using CAPCOST, we have evaluated the profitability of the simulated process. With a positive net present value and an attractive discounted internal rate of return, the simulated plant stands out with a short payback period of 3.5 years. Thus, overall, our simulations also reasonably represent the economic feasibility of the CWM process. This appealing outcome is encouraging for devoting more efforts to study modern process integration options in Aspen Plus environment in the future.