Chapter 7: General Discussions
7.1 Stand-alone systems meeting electric load only
When using single objective Genetic Algorithms (GAs) to optimally size systems satisfying highly dynamic electric load, the study first analyses the effects of several key factors on the Cost of Energy, reliability of meeting demand, renewable penetration, duty factors, and environmental impact (Chapter
2). The study also includes the optimisation of a hybrid system using HOMER with different battery
technology (Chapter 6). In the following discussion, the issues related to research questions RQ 1 and RQ 5 have been addressed.
• Cost of Energy: The results of this study show in Chapter 2 that the type of prime movers (i.e. ICE or MGT) and their configurations have significant effect on Cost of Energy (Table 2.4). The cost for renewably-based system with battery storage is comparable with the ICE-based hybrid energy systems regardless of a single large unit or multiple units used in tandem. However, the COE for ICE-based hybrid system is considerably lower than the MGT-based system. From the sensitivity analysis it can be said that the changes of interest rate and overall capital costs of components have noticeable effect on the COE for both scenarios while the other parameters (fuel cost, PV costs, and capital costs of ICE) have the least effect (Figure 2.9).
• Performance indicators: Combustion based hybridised systems produce significant amounts of waste heat as well as CO2 (operational) and life cycle emissions, even though the COE may
175 | P a g e
be similar to PV/Batt systems. This is likely to be exacerbated if a thermal load (through cogeneration or trigeneration) also needs to be met, which then impacts overall efficiency not just COE. The results of this study indicate that the ICE-based system is more environmentally sustainable than the MGT-based hybrid system (Table 2.4). Although, the COE for the MGT based hybrid system is much higher than the hybrid ICE system, the MGT produces 316% waste heat to the electric power supplied by it. This figure is considerably lower for the ICE (129% waste heat to supplied electric power) and could possibly make the MGT system a more favourable supplementary device for cogeneration applications (Table 2.4).
• Hardware parameters: The COE for system consisting single unit supplementary prime mover is comparable to the similar capacity multiple prime movers operating in tandem (Table 2.4). However, the waste heat generation from the multiple units of supplementary prime movers is significantly higher than a single capacity engine at the cost of LCE. This is because of lower starting threshold constraints of ICEs or MGTs, which force to start engines when the load demand is lower (i.e. 9kW for 30kW engine) than the larger single capacity engine (i.e.18kW for 60-65kW). The transient start-up of supplementary prime movers have insignificant effects on sizing for both PV/Batt/ICE and PV/Batt/MGT-based hybrid systems, which account missed load 35kWh and 210kWh, respectively, over the year. The loss of load demand due to the transient behaviour of prime movers are well below the loss of power supply constraints (760kWh, LPSP: 0.01±0.005) considered in this analysis for system sizing. Although the minimum starting threshold (Psup, min) of supplementary prime movers has no
considerable effects (4–7%) on COE, it does on the LCE and operational CO2 emissions (Figure
2.7) because of the higher running at lower Psup, min. From this study in Chapter 2 indicate that
the effects of Psup, min for the system involves ICE is less pronounce in relation to the waste heat
generation to the supplied power. However, this effect is more significant for MGT-based hybrid system that ICE-based system (Figure 2.7 (d)). This is due to the engine characteristics that produce more heat to power for MGT than the ICE.
• Optimisation methods: The study presented above is carried out using population size 10 as there appears no appreciable further improvement of objective function (Cost of Energy) but the time is considerable higher at larger population size than 10 (Appendix E). The effects of temporal resolution (60min vs 15min) of meteorological (i.e. solar irradiation, wind velocity, and ambient temperature) and electric load data have been discussed in Chapter 2 (Figure 2.8). The results from this indicate that this effect is insignificant on the COE and the waste heat generation. However, there are noticeable effects on emissions for both PV/Batt/ICE and PV/Batt/MGT-based hybrid systems which lead to further investigation of this study using 15min resolution for CHP (Chapter 3) and CCHP (Chapters 4 and 5) systems. Chapter 6 investigates a stand-alone PV/ICE system with different battery technologies (i.e. Lead acid, Li-ion, Vanadium redox flow) both economically and environmentally. Optimisation of this
176 | P a g e
study reveals that both Lead Acid Battery (LAB) and Li-ion-based hybridised system has comparable COE, whereas, the system involves Vanadium Redox Flow (VRF) battery cost is relatively higher because of the higher capital cost of battery (Table 6.5). Although capital and annualised costs associated with PV/ICE/Li-ion and PV/ICE/VRF-based systems are higher than the PV/ICE/LAB-based systems, the replacement and O & M costs are considerably lower in systems involves in Li-ion and VRF than the LAB (Tables 6.5 and 6.6). Although the COE and NPC for the PV/ICE/Li-ion is comparable to the PV/ICE/LAB, the affordability (lower battery capital cost) availability across the world, and extensive study of using lead acid battery (Table 6.2) encourages further study with lead acid battery.