(554c) The Use of Cross-Interval Transshipment (CIT) Model to Optimize Power Cogeneration: The Case of Multiple-Product Biorefineries

Chemical engineering faces with candidate chemistries and feedstocks towards the production of biochemicals; all candidates are present in value chains and need to be screened and appropriately selected to secure sustainable establishment of bio-economy applications. Chemistries are regularly tested in industrial biotechnology and circular economy; however, high energy costs are often unable to support sustainable introduction of selected chemistries and bioproducts, unless proper and strong integration is provided. This paper presents an updated version of a recently demonstrated integration tool (Pyrgakis and Kokossis, 2019), which is used for the simultaneous synthesis and integration of biorefineries.

The integration tool searches for cost-effective and complementary value chain chemistries that optimally share energy and materials building high efficient biorefinery applications. Beyond the appropriate selection of bioproducts, power cogeneration (using steam turbines) still counts as an additional and valuable co-product necessary to offset price volatilities of upcoming biochemicals markets. Turbine efficiencies and power production are strongly affected by the selected biorefinery chemistries and the integration patterns among them. In this scope, screening of value chain processes should be examined in view of both energy savings and cogeneration targets, at the same time, to minimize processing costs and boost electric power production.

Scoping for optimal efficiencies, steam savings from process-to-process integration (via steam generation ad reuse) naturally competes with power cogeneration. Once steam is generated by available process heat, challenges arise about the distribution of generated steam either to steam turbines maximizing cogeneration and using the effluent (lower pressure) steam for integration (losing integration efficiencies) or directly for process-to-process integration maximizing steam savings (but losing power cogeneration) or share to both routes searching for maximum economic performance. As far as processes and utility levels of the understudied biorefinery – or any multiple-process Total Site – are given beforehand, the estimation of energy targets is straightforward by means of robust optimization. In other words, given the site utility demands and the selected utility levels, the implementation of Turbine Hardware Model (THM) of Mavromatis and Kokossis (1998) is enough to maximize cogeneration performance and address trade-offs between steam savings. This has been discussed by Shang and Kokossis (2004) with the use of a tree-cascade representation combined with the THM. However, in the case of biorefineries, where candidate processes can be combined and get integrated by any means, the size and complexity of integration-cogeneration problem explodes. The conventional graphical and computational tools are strictly limited for known and given processes.

One should systematize process-to-process integration– that used to be exclusively provided by means of the graphical Site Sources and Sinks Profile (SSSP) – and next combine with cogeneration technologies to drive the overall optimization problem. Such a solution has been provided in a previous work of Pyrgakis and Kokossis (2019) in the form of a Cross-Interval Transshipment (CIT).

The CIT constitutes an extended heat transshipment approach, which holds additional nodes and heat flows to model indirect integration via steam along with direct integration, at the same time. Besides the conventional energy nodes that are employed for debottlenecking each (individual) involved process, CIT implements additional external nodes (Pyrgakis and Kokossis, 2019) that supervise heat flows and available heat across all temperature intervals. The nodes are capable to collect heat from higher intervals; translate heat into generated steam; and serve steam to lower temperature intervals, finally connecting non-consecutive intervals and providing indirect integration. At the same time, the nodes inside intervals are still responsible for direct integration. As a result, the CIT simultaneously models direct and indirect integration and mathematically formulates the graphical procedure (SSSP). The implementation of CIT introduces a new heat cascade concept – the Total Site Cascade (TSC) – that is available for integration of single and especially multiple-process Sites.

Thanks to the additional energy nodes, the TSC is also capable to supervise, manage and optimize both fresh and generated steam loads along the entire temperature range of the biorefinery. These nodes constitute an ideal solution to monitor the available steam loads at any steam level (even for a set of candidate utility levels) and to estimate cogeneration potentials for any candidate expansion zone; namely, the span between two steam levels. In addition, the nodes reflect to each utility level operated by the Total Site and the user is free to use multiple candidate utility levels (e.g. CIT nodes) to proceed in optimization of steam levels. The TSC holds extra options to model the heat contribution of streams considering processes as additional degrees of freedom. The processes can be selected by any available superstructure, like the Biomass Bipartite graph Representation (BBR) – also presented by Pyrgakis and Kokossis, 2019 – which maps all feasible process synthesis options along candidate processes. In this work, CIT is extended with cogeneration technologies (THM) to approximate turbines power output and new strategies to improve accuracy of cogeneration models.

Given a set of candidate biorefinery chemistries to select and integrate, the CIT model is employed to maximize integration and cogeneration efficiencies. A dense grid of multiple, candidate utility levels are also considered to explore available heat along the entire temperature range – configured by all candidate streams – and appropriately place steam levels so as to optimize integration and cogeneration efficiencies.

The proposed methodology takes advantage of the Turbine Hardware Model (THM) (Mavromatis and Kokossis, 1998) to provide good approximations of shaft-work production. Complex turbines are decomposed into single backpressure turbines that receive/extract steam at multiple expansion zones eliminating non-linearities of turbine efficiency with load. This work addresses heat sources/sinks and utilities as degrees of freedom, while integration takes place over the TSC, which hosts heat flows to model direct-indirect integration and cogeneration at the same time.

BBR is used to formulate mass balances along value chains, to generate all biorefinery site combinations and to manage their heat contribution to the TSC. Mass (BBR) and energy (TSC) balances are constructed as an optimization model (MILP) to select processes, products and utility levels that maximize the biorefinery profitability considering energy cost, opex-capex, power cogeneration and products profits. The updated model is illustrated through a real-life biorefinery case succeeding in high steam cost savings (20%) and power cogeneration; the latter counts more than one third of biorefinery profitability.

References

1.Pyrgakis, K., Kokossis, A., 2019. A total site synthesis approach for the selection integration and planning of multiple-feedstock biorefineries. C&CE, 122, 326-355

2. Mavromatis SP, Kokossis AC. Conceptual optimisation of utility networks for operational variations-II. Network development and optimisation. Chemical Engineering Science. 1998; 53;8:1609

3. Shang, Z., Kokossis, A., A transhipment model for the optimisation of steam levels of total site utility system for multiperiod operation. Computers and Chemical Engineering, 28, 1673-1688