Seite - 91 - in Short-Term Load Forecasting by Artificial Intelligent Technologies

Energies2019,12, 57 the proposed iterative optimization canbe applied for a quick estimation of the systembehavior. Simultaneousoptimizationcanbeconsideredforpotentiallybetteroptimization if timepermits. 3.2. LaboratoryLoad Similarly, theresultsofapplying the laboratory load(seeFigure5) to thehybridpowermodel are shown in Figure 7 and Table 6. First, the original system layout (b, s, w) = (8, 10, 2) with managementsettingsof (SOClow,SOChigh)= (40%,50%)resulted inasystemcostof J=1.100$/kWh and LPSP=26.42%. Note that the LSPSwasmuch higher thanwas obtained for the household, because the laboratory loadwasmainlyatnightandthestoredenergybyhydrogenelectrolyzation failed toprovide sufficient energy. The initial componentoptimizationcan reduce the systemcost to J=0.929$/kWhby setting (b, s,w) = (27, 15, 2) butwith LPSP= 2.34% (see Step 2 of Table 6). Thesub-optimal settingsof (b, s,w)= (30, 16, 2)gaveLPSP=0with J=0.944$/kWh(seeStep3of Table6), i.e., thereliabilitywas improvedby26.42%,while thecostwasreducedby14.18%. ȱ (a)ȱwȱ=ȱ2ȱandȱ(SOClow,ȱSOChigh)ȱ=ȱ(40%,ȱ50%)ȱ (b)ȱwȱ=ȱ0ȱandȱ(SOClow,ȱSOChigh)ȱ=ȱ(30%,ȱ40%)ȱ Figure7.Thereferenceplots for the lab load. Table6.Theoptimaldesignprocedures for the lab load. (b,s,w) (SOClow,SOChigh) LPSP (%) J ($/kWh) Step1 (8,10,2) (40%,50%) 26.42% 1.100 Step2 (27,15,2) (40%,50%) 2.34% 0.929 Step3 (30,16,2) (40%,50%) 0% 0.944 Step4 (31,21,0) (40%,50%) 0% 0.684 Step5 (31,21,0) (30%,40%) 0% 0.668 Step6 (27,21,0) (30%,40%) 0% 0.660 Step7 (27,21,0) (30%,40%) 0% 0.660 Optimal (27,21,0) (30%,40%) 0% 0.660 Because theWTwasnoteconomicallyefficient for thisbuilding, settingw=0cangreatlyreduce the systemcost to J=0.684 $/kWhwithLPSP=0by (b, s,w) = (31, 21, 0) (see Step 4 of Table 6). The iterativeprocedurescould thenfurther improve thesystemcost to J=0.668$/kWhwithLPSP=0 bysettingthepowermanagementas (SOClow,SOChigh)= (30%,40%),andthecostfinallyconverged to J=0.660$/kWhwithLPSP=0bysetting(b, s,w)= (27,21,0)and(SOClow,SOChigh)= (30%,40%). Whencomparedwith theoriginal cost, thecostwasreducedby40%,while thesystemreliabilitywas reducedby26.42%. 3.3.OfficeLoad Theanalysesof theoffice load(seeFigure5)areshowninFigure8andTable7. First, theoriginal systemlayout (b, s,w)= (8,10,2)withmanagementsettingsof (SOClow,SOChigh)= (40%,50%)gavea systemcostof J=1.107$/kWhandLPSP=5.08%.Optimizingthesettingsslightlyreducedthesystem cost to J=1.106$/kWhwithLPSP=0using(b, s,w)= (23,11,2) (seeStep2ofTable7).Note that the 91