First generation surrogate models for sustainability assessment of biobased chemical production and integrated biorefineries
Stavros Papadokonstantakis. From: Swiss Federal Institute of Technology (ETH), Zurich, to: Chalmers University of Technology.
Surrogate models are simple algebraic meta-models with a multi-functional, explanatory and exploratory role in the substitution of rigorous, yet often cumbersome to use in decision-making and optimization, process models to a required accuracy and revealing the underlying functional relationships between process parameters and key performance indicators. This project aims at a first generation of surrogate models for sustainability assessment in early and conceptual design stages of biomass based chemical production and integrated biorefineries. The models will utilize available information at the respective design stages, such as molecular descriptors and physical properties of the compound or mixture of compounds produced in dedicated or integrated biorefineries, generic chemical and process synthesis information, the degree of heat and mass integration, the level of process flexibility and the technological maturity to provide estimations about performance indicators with respect to process costing, LCA and SHE hazard assessment. The surrogate models will be developed on the basis of available rigorous process models and data from synergies with running projects in Chalmers (BioBUF, financed by FORMAS) and ETHZ (Wood2Chem, financed by the Swiss National Science Foundation) as well as various partners and projects in the field of biomass utilization (e.g., BIOCORE (FP7, NTUA), RENESENG (FP7, NTUA, DTU)). The project will also provide estimations for a wide list of indicators from the GREENSOPE taxonomy (US-EPA). The surrogate models will support policy decision makers, LCA practitioners, process engineers and decision-makers in industry and will also contribute to a guidance of R&D activities between industry and academia to the most promising technology directions in biomass utilization from a holistic sustainability point of view.
Project Summary Results:
At local, regional and global levels, there is an increasing demand for chemical industry to provide products, services and innovations in a way that is benign towards environment and society. To meet this challenge, an overall vision is required to develop innovative, trans-disciplinary approaches to integrate modeling, assessment and optimization of chemical processes and products during the whole life-cycle of a chemical, including all three pillars of sustainability, namely environmental and socio-economic ones. This project contributed to this vision in two ways. First, databases of biomass-to-product process chains were constructed following a modular approach for upstream and downstream unit operations in these process chains. These databases include a thorough list of environmental impact assessment indicators (i.e., mainly from a life cycle perspective) and cost factors. Then, based on these databases, surrogate models of various types were developed (i.e., based on regression and classification techniques) for prediction and generalizing the performance of biomass based production in terms of cost, integration potential and environmental impact.
This approach facilitates a transparent environmental life cycle impact assessment for bio-based products. It also enhances the ability to develop and assess more complex biorefinery systems, identifies critical parameters and offers useful material to support decision making in early stages. The databases and surrogate models consist of more than 100 study systems, including various forms of feedstock types (e.g., wood chips, miscanths, vegetable oil), undergoing transformation to various products of interest for the chemical industry (e.g., organic acids, alcohols, polymers) and the fuel sector (e.g., biodiesel, biogas, gasoline).
The project was conducted in close collaboration with ETHZ (i.e., in particular utilizing data from the Wood2Chem project, financed by the Swiss National Science Foundation) and NTUA (i.e., in particular utilizing data from BIOCORE and RENESENG, financed by EU-FP/). It is envisaged that the databases and surrogate models will provide LCA practitioners, process engineers and decision-makers in industry a source of data, tools as well as a benchmark for assessment of individual projects and thus a guidance of R&D activities between industry and academia to the most promising technology directions in biomass utilization from a holistic sustainability point of view.