Sizing of a cogeneration unit based on fuel cells and on steam reforming of diluted bioethanol

Introduction The production of electric energy for small-scale applications, using hydrogen as energy vector, can rely nowadays on the already established fuel cell technology coupled to efficient reforming catalysts and processes [1-3], that set renewable energy sources as viable alternatives to the traditional energy production and distribution chain. Bioethanol, among other feedstocks, is very promising to meet the demand of the large market of distributed energy cogeneration. With respect to reforming plants based on pure ethanol, different strategies can rely on less expensive diluted bioethanol. In this case, the system is much more flexible both for the fuel supply and for the possible operation modes, provided that the additional heat input required to vaporize excess water quantities can eventually be retrieved as useful thermal energy. With this work, both steady state and dynamic simulation were used for a detailed design of a residential size cogeneration system based on a bioethanol-to-hydrogen-to-power technology. The basic layout and preliminary sizing of a water circuit has been also performed, that recovers a substantial fraction of the waste heat released by a bioethanol reforming unit already sized to meet a home-scale power production of 5 kWelectric. The thermal energy is sufficient to cover the steady-state wintertime dispersion for a class ‘F’ two-storey house in the Northern Italy climate with a traditional radiation system, while the more demanding sanitary water (DHW) production can be met by resorting to a micro-accumulation strategy. The flexibility provided by a diluted hydro-alcoholic feed mixture with a separate reformer-fuel cell layout makes it possible to tune the system performance according to the instant house need, and the use of hydrogen as a vector rather than as a source let foresee an easier feedstock supply and management. Models and methods The calculation of the reforming apparatus has been performed using Aspen Plus® V8.8 with the PURE32 Databank. The Peng-Robinson and NRTL thermodynamic systems have been employed for the whole flowsheet, while for the simpler steady-state radiators power we adopted the STEAM-TAB plus PENG-ROB packages. The dynamic simulation of the DHW (Hot Water) delivery has been modeled and solved with Matlab® V7.10 (using the ‘ode45’ algorithm to integrate the differential equations). The algorithm used to estimate a model house heat dispersions runs on MS Excel™. Results and discussion An increased water content in the feed, i.e. the use of diluted bioethanol, has an overall beneficial effect on the steam reforming system for hydrogen production. On one hand, the heat subtracted to the burned gas downstream the reformer (Fig.1), once the hydrogen production is accomplished, is released anyway at the condenser. Furthermore, the CO purification is easier through waster gas shift and, hence, the methanation step can be downsized. The lower temperatures and flowrates at the burner exit makes the reformer operation much more stable with respect to the gases split fraction, eliminating the instability range which prevents the increase of the FC power at intermediate utilization factors. A high water dilution, on the other hand, is not compatible with a full FC exploitation of the reformate, needing separation in a condenser. High dilution can give raise to tricky regime-change and instability phenomena at too low values of the utilization factors of the fuel cell, where the hot gases production and utilization are both enhanced to opposite effects. The details of this behavior are anyway dependent on the absolute ethanol quantity employed, since this value fixes the total power available from the system in any form (electrical or thermal), but we can conclude that a ratio of 7 moles of water to 1 mole of alcohol is preferable to a more concentrated 5:1 feed for this power scale. The increase of the condenser duty (insensitive to the flue gases split fraction) at the burner gas expenses, makes the total thermal power available at lower temperatures. Our steady-state calculation of a typical radiator-based house heating network shows that there are sufficient operation margins to overcome this limitation even in the worst winter case, bringing the overall system energetic yield to a minimum of 52%, over a theoretical 80%. The sanitary water production scenario is different (Fig.2) and too high condenser duty makes unfeasible reaching the set-point. In all the other cases, however, the hot water demand of 10 L/min can be met essentially in two ways: i) decreasing the electrical output to 3 kW for 15 minutes while maintaining the water content at its nominal level or ii) maintaining a high electrical power output and letting the sanitary circuit deplete for as long as 35 minutes. According to different household electrical needs and environmental situations (winter / summer), the switch between this strategy is readily achievable, since the explored best cases for the hot water supply cover both the cases of ethanol dilution with 5 or 7 moles of water at different fuel cell power. Moreover, this calculation is quite conservative since it is supposed that more than half of a family daily consumption is consumed in 30 – 40 minutes, only. Moreover, when the sanitary water is being produced, the plant efficiency reaches its maximum ideal value, because all the heat dissipated from the reformer goes directly (via the sanitary HX) or indirectly (via the main water reserve) into this circuit. If the condenser temperature is raised to 55 °C, the cogeneration section gains two advantages: i) the steady-state radiators calculation shows an increased thermal recovery (nearly 1 kW of the 21 globally available), and ii) the sanitary water circuit can be set to work at no less than 48 °C for at least 35 minutes without sacrificing the electric production. In this context, if a lower hot water target temperature is maintained, it is also possible to shift the heat balance from the hot gases to the condenser, increasing further the overall efficiency to 57%. References [1] V. Kirillov, V.D. Meshcheryakov, V. Sobyanin, V.D. Belyaev, Y.I. Amosov, N. Kuzin, et al., Theor. Found. Chem. Eng., 42 (2008) 1–11. [2] I. Rossetti, M. Compagnoni, M. Torli, Chem. Eng. J. 281 (2015) 1036–44. [3] I. Rossetti, C. Biffi, G.F. Tantardini, M. Raimondi, E. Vitto, D. Alberti, 37 (2012) 8499–504.