The interplay between renewables penetration, costing and emissions in the sizing of stand-alone hydrogen systems
Faculty of Health, Engineering and Science
School of Engineering
Multi-objective Genetic Algorithms are used to optimise three stand-alone hydrogen systems (WG-H2, WG/PV-H2 and PV-H2) under three different objective functions: minimising (hardware) Net Present Cost - NPC ($), whole Life Cycle Emissions - LCE (CO2-eq/yr) and dumped/Excess Energy -EE (%) at low demand. Optimisations considering Excess Energy haven't been reported before. Simulations are implemented using MATLAB, incorporate experimentally resolved fuel cell start-up transients, and dynamic profiles for wind speed, solar irradiance as well as electric load demand. Results indicate the significance of integrating fuel cell start-up into the LPSP when optimising systems, another aspect not reported before and a modified LPSP is introduced. Furthermore, when sizing energy systems by reducing LCE, EE, and NPC, the favoured hybrid architecture appears to be WG-H2 over the others studied. For the same LPSP, an interesting finding is that increased renewables penetration (reduced dumped loads) affects the optimised solution but comes at a cost.