Seite - 85 - in Short-Term Load Forecasting by Artificial Intelligent Technologies

Energies2019,12, 57 ȱ Figure1.Thegreenbuilding. Table1.Thedailyaverageweatherdataonthebuildingsite [32,34]. DataSource Irradiance (W/m2) WindSpeed(m/s) Summer Winter Summer Winter NASA[34] 267 109 4.95 8.70 Measured 239 115 2.42 3.96 Table2.Energycostanalyses ($/kWh)[32]. EnergySources Summer Winter Photovoltaic (PV)arrays 0.11 0.23 Windturbine (WT) 7.76 0.69 Protonexchangemembrane fuel cell (PEMFC)withchemicalH2generation 1.76 1.76 2.1. TheHybridPowerModel A general hybrid power model, as shown in Figure 2, was developed to evaluate system performance at different operating conditions (e.g., varying the component sizes and power management strategies) [32,33,39]. Themodel consistedof aPVmodule, aWTmodule, a battery module,aPEMFCmodule,anelectrolyzermodule,achemicalhydrogengenerationmodule,anda loadmodule. Thepowermanagement strategieswereapplied tooperate thesemodulesbasedon batterystate-of-charge (SOC).Themoduleparameterswereadjustedbythecomponentcharacteristics andexperimental responses toallowpredictionandanalysisof thesystemdynamicswithout theneed forextensiveexperiments [39,40]. First, the1kWPVmodulewasdevelopedbasedonthe followingequation[32,41]: PPV=0.69(E−1.52) (1) wherePPV (Watt)andE (Wattpermetersquare) representsolarpowerandirradiance, respectively. Second, theWTmodulewaspresentedasa look-up table, according to the relationbetweenwind power andwind speed [33,42]. Third, thePEMFCacted as a back-uppower source to guarantee systemsustainabilitybasedonthe followingmanagementstrategies (seeFigure3a) [39]: (1) When thebatterySOCdropped to the lowerbound,SOClow, thePEMFCwas switchedon to provideadefault currentof20Aat thehighestenergyefficiency[41]. 85