(534f) Total Site Analysis As a Synthesis Model to Select, Optimize and Integrate Processes in Biorefineries

The paper presents a new approach to apply energy integration in multiple-product plants. The incentives of this work originate in biorefineries. Biorefineries constitute the most promising route for the valorisation of bio-renewables and the replacement of fossil-oil in any application of chemicals and fuels production. Biorefineries are capable to utilize a wide range of biomass sources including agricultural and forestry residues, industrial and municipal wastes as well as aquatic biomass (algae). Several biorefineries have already taken off in USA, EU, China and Brazil; the majority of applications focus on the production of bio-fuels. However, a big number of conventional and novel bio-products â?? bulk chemicals and specialties, polymers, coatings, resins, pharmaceuticals, enzymes, proteins, fuels etc â?? can be generated via thermo-processing and bio-processing of biomass. Bio-chemicals production is projected to reach up to 10 % of $1.5 trillion worldwide annual chemicals market [1], while biorefineries are capable to consume up to 84 % less non-renewable sources increasing expectations for GHG emissions reduction from 60 % to 100 % [2].

Biorefineries can be benefited by the integration of multiple products and production lines through the exchange of materials and energy among them. The development of appropriate synergies among the different processes may result in significant reduction up to 90 % in CAPEX [2], 84 % in energy consumption [3] and 58 % in water demands [4]. Provided that numerous products and technologies can get integrated in biorefineries, critical questions rise about which processes portfolio maximizes efficiencies, if collocated and integrated together. Given the high impact of energy cost in sustainability of biorefining, this work focuses on energy integration techniques in upcoming applications.

Energy integration is a powerful tool that estimates the maximum feasible margins of energy savings by exchanging heat between hot (need cooling) and cold (need heating) streams involved in the process (direct integration). In the scope of integrating multiple-process plants (like biorefineries), energy integration takes the form of Total Site Analysis (TSA) [5]. TSA estimates steam utility savings by using excess heat from the individually (directly) integrated processes to cover steam demands of others via steam production-reuse (indirect integration). The graphical tool of Site Sources and Sinks Profiles (SSSPs) is used to estimate steam savings in Total Sites. However, the graphical tool appears impractical and ineffective to address the upcoming industry, since assumes fixed and known processes. In the case of biorefineries, processes are not essentially known and solutions are not profound given the range of candidate technologies that may enter the biorefinery. The biorefinery problem constitutes a synthesis problem, where processes are significant degrees of freedom in integration. Each process features a different energy profile and energy savings will change as far as processes are integrated across different process combinations. In contrast to conventional TSA, heat exchange options in biorefinery cases do not refer to a fixed cluster of processes, but actually expand among all candidates. A potential solution would be the exhaustive use of the graphical tool (SSSPs) to estimate steam savings for all possible combinations and detect the one with the lowest energy cost. In fact, this approach is impractical for large problems (number of combinations explodes) and appears to be inefficient and unsafe to solve the biorefinery problem, as it is proven in this work.

The candidate biorefinery routes from raw materials (biomass) to intermediate and final products are represented in the form of value chain paths. Each path denotes a candidate process, while a set of consecutive paths composes a potential biorefinery plant (Total Site). Alternative biorefinery structures can be developed by choosing each time different processes among competitive paths (they valorise the same chemical), which are branched along the value chain. In the scope of energy integration, the challenge is to systematically address all alternative structures and detect the biorefinery that maximizes steam savings and minimizes energy cost. This means that screening of process combinations and integration (direct and indirect) should simultaneously occur. To address the complexity and multiplicity of the synthesis problem, Total Site Analysis should be upgraded into a synthesis tool by combining thermodynamics with mathematical programming.

The use of SSSPs requires fixed and known processes and, thus, known energy contents of streams that will get integrated. This implies that processes from the value chain should be previously selected and get integrated using SSSPs; accordingly, one process among competitive can be selected and integrated with others each time. Nevertheless, higher savings can be succeeded by sharing upstream chemicals among competitive processes in any way, subsequently changing the capacity and the energy content of involved hot/cold streams in order to improve the heat exchange conditions and secure better synergies. In the courses of a real-life biorefinery problem, this approach resulted up to 8 % higher steam savings compared with savings obtained by using SSSPs. Provided that the graphical tool is not capable to meet with these needs, new comprehensive representations are proposed to aid with the mathematical formulation of Total Site integration and address process capacities as additional continuous variables. The proposed concept could be interpreted as if SSSPs involve moving (variable) curves.

Two new representations, which communicate each other, are introduced to describe the combinatorial synthesis and integration problem. The first, a Biomass Bipartite graph Representation (BBR) is used to translate value chain paths into representations that enable the development of mass balances along the value chain. The second, a Total Site Representation (TSR) uses a properly modified transshipment model that incorporates all heat exchange options (energy balances) related with direct and indirect integration and enable the mathematical formulation of Total Site integration. The mass and energy balances are constructed as an optimization model that is solved using mixed integer linear programming and yields the processes portfolio that minimizes energy cost of the under-construction plant.

The BBR uses three building blocks: (i) product nodes for chemicals (raw, intermediate and final), (ii) process units for paths and (iii) interconnections that connect nodes with units and describe the transportation of chemicals from path to path. The flowrate of each chemical is assigned to the relative interconnection and, thus, mass balances can be recorded for each product node and process unit, given the capacity of raw materials at the beginning of value chain and the conversion rate of each path (given by experimental groups). Accordingly, the flowrates of process inlet interconnections are used to denote the capacity of the associated process. Each path is previously simulated to estimate the normalized heat demands of involved streams per unit mass of inlet chemical. Assuming that mass and energy balances within a process are linearly depended on its capacity, the actual heat content of hot/cold streams can be linearly expressed by the normalized heat demands and the variable flowrates involved in mass balances. This expression secures the connectivity between BBR (selection of paths) and TSR (integration of selected paths).

The TSR incorporates a super-interval cascade, called Total Site Cascade (TSC), which is composed by the temperature intervals of all candidate processes. The process streams contribute to TSC according to flowrates selected by BBR and the modified cascade proceeds with the direct integration of all selected processes simultaneously rather than each separately. As a result, the utility needs of individually integrated processes are estimated by TSC. At the same time, a parallel set of blocks, called utility operators, gathers the utility needs from TSC and proceeds with indirect integration. Specifically, the utility operators (i) translate excess heat from TSC into steam production, (ii) estimate the overlap between generated and demanded steam and (iii) call for external (Total Site) utilities to cover the remaining demands. In fact, the utility operators identify which process are capable to work as steam-source or sink and detect (in combination with BBR) energy complementary process combinations. The energy balances at each internal of TSC and the utility operators are constructed as an optimization model that minimizes energy cost of Total Sites. The combined system of TSC and the utility operators composes the TSR. The model composed by the energy balances of TSR can be also individually used (without BBR) for Total Site integration of a fixed cluster of processes avoiding the graphical tool of SSSPs.

The proposed methodology was applied for the development of a real-life lignocellulosic biorefinery with 9 candidate paths and 11 bio-products yielding in the 5 best processes portfolio that consume 9 % and 14 % lower hot and cold utilities, respectively, compared with the best solution that obtained by exhaustively using SSSPs.

Literature

1. Austin A., Biomass Magazine, 2012, http://biomassmagazine.com/articles/2778/bio-roundtable-feedstock-flexib...
2. EuropaBio, 2011, http://www.europabio.org/biorefinery-feasibility-study
3. Kokossis A.C., Tsakalova M., Pyrgakis K., 2015, Design of Integrated Biorefineries, Computers & Chemical Engineering, 81, p.p. 40â??56
4. Nikolakopoulos A., Karagiannakis P., Galanis A., Kokossis A., 2012, A water saving methodology for the efficient development of biorefineries, Computer Aided Chemical Engineering, 30, p.p. 7-10
5. Dhole V.R., Linnhoff B., 1983, Total Site Targets For Fuel, Co-Generation, Emissions, And Cooling, Computers & Chemical Engineering, 1, S101-S109