(71d) A Systematic Approach for the Characterization of Non-Conventional Streams in Biorefinery Applications

Biorefinery plants convert biomass feedstock to valuable biobased materials like chemicals and fuels, offering a green alternative to fossil fuel consumption. Biorefineries are classified by feedstock type, processing techniques, platforms and products, hence allowing for process flexibility and product diversification. Figure 1 illustrates integration of three biorefinery platforms, visually explaining the adaptability of biorefinery systems. Appropriate integration of biorefinery systems to industrial facilities would significantly contribute to the sustainable production of valuable chemicals, materials and fuels, while ensuring energy security and market opportunities. The significance of biorefinery plants is undeniable despite the fundamental challenges that need to be addressed. Proper substrate selection exhibits high complexity due to the variability of biomass feedstocks. Biomass feedstocks are inherently diverse, comprising of various organic compounds depending on availability, source, geographic location and seasonal conditions. The diversity of the substrate leads to significant variations in the chemical composition, thermodynamic and physiochemical properties of the produced biomaterials. Flowheeting techniques are available for proper process synthesis but are often inadequate due to the lack of thermodynamic knowledge associated with non conventional materials. Hence, it is essential to develop standardized techniques for feedstock and product profiling. A highly promising biorefinery technology, illustrated in Figure 2, for the production of valuable biofuels and biochemicals from organic waste is Hydrothermal Liquefaction (HTL). HTL is a continuous process which can accommodate a wide range of feedstocks (solid, liquid or even sludge) regardless of their moisture content, thereby eliminating the need for energy intensive waste pre-treatment (e.g. drying). Essentially, HTL offers high adaptability and resource efficiency rendering it a viable option for biomass utilization.

This paper presents a generic mathematical programming approach formulated and solved using MILP and MINLP technologies to efficiently characterize both biomass feedstock and biocrude streams with a focus on HTL. The problem is formulated with:

• Variables that include

o Chemicals in available databases

o Composition in the solution

• Specification parameters that icnlude

o Substrate classification (e.g. breakdown of proteins, sugars, lipids etc.)

o Thermodynamic properties (e.g. densities, HHV, HLV, viscosity)

o Elemental and stoichiometric composition (e.g. ratios of C:H:O:N:P)

o Experimental measurements (e.g. moisture content, fixed carbon)

• Objective functions featuring

o a vector stream with suitable matching properties and relevance to the nature of the substrate

The problem is formulated as MINLP with constraints that include property constraints; logical constraints that relate to the selection of chemicals , and integer cuts to produce populations of solutions that are further analyzed and clustered. The model has been tested on 5 classes of substrates with results that indicate a wide range of multiple solutions and a rather economical representation as compared with other work in literature (75% reduction with similar performance). Figure 3 illustrates the prediction of elemental and biochemical composition of a substrate for a number of chemicals in comparison to literature models and experimental values. The presented formulation produces elemental and biochemical composition close to the experimental values, proving its efficiency. The model’s economical representation is proven when compared to the one from literature, where a 75% reduction of the chemicals proposed in the literature and better results. Figure 4 illustrates how the objective function fluctuates with respect to the number of chemicals present in the substrate. The objective function value is higher at the beginning, which corresponds to a small number of compounds. As the compound number increases (>5) the objective function value remains close to 0 indicating an efficient feasible solution.

Overall, the model succeeds in (i) estimating key properties by 75% reduction in required chemicals, thus facilitating digital representation of material streams (ii) calculating key thermodynamic properties (e.g. density, HHV) (iii) evaluating a number of feasible solutions to determine the best one. The results of the formulated model were documented for 5 different substrates and suggest that biomass profiling can be efficiently and economically achieved. The paper produces a viable optimization model which efficiently establishes key properties of different substrates, hence providing significant insight in biomass profiling. The approach is being extended to involve all the chemical compounds in a database and to point to classes of solutions that deliver acceptable results (using integer cuts). Results produce data pools where chemicals can be decomposed using group contribution methods. Groups can be further used as input to machine learning algorithms to propose representations from groups (rather than chemicals).

Acknowledgements

The project is implemented within the framework of the National Recovery and Resilience Plan "Greece 2.0" and has been financed by the European Union (NextGeneration EU).

References

1. Vijai Kumar Gupta, Tuohy, M.G., Kubicek, C.P., Saddler, J.N. and Xu, F. (2014).

Bioenergy Research: Advances and Applications. pp.219–241. doi:https://doi.org/10.1016/c2012-0-00025-7.

2. Watson, J., Wang, T., Si, B., Chen, W.-T., Aierzhati, A. and Zhang, Y. (2020). Valorization of hydrothermal liquefaction aqueous phase: pathways towards commercial viability. Progress in Energy and Combustion Science, 77, p.100819. doi:https://doi.org/10.1016/j.pecs.2019.100819.

3. Aslanoglou, I. (2022). Modelling of non-conventional streams in circular economyAnalysis of HTL streams. doi:http://dx.doi.org/10.26240/heal.ntua.24046.

4. Gollakota, A.R.K., Kishore, N. and Gu, S. (2018). A review on hydrothermal liquefaction of biomass. Renewable and Sustainable Energy Reviews, 81, pp.13781392. doi:https://doi.org/10.1016/j.rser.2017.05.178.

5. Krushna Prasad Shadangi, Prakash Kumar Sarangi and Amit Behera (2023).

Characterization techniques of biomass. Elsevier eBooks, pp.51–66. doi:https://doi.org/10.1016/b978-0-323-98363-1.00022-3.