Risk due to load forecast uncertainty in short term power system planning

Douglas, A.P.; Breipohl, A.M.; Lee, F.N.; Adapa, R.

doi:10.1109/59.736296

Cited by 92 publications

(39 citation statements)

References 5 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The references [29,30] use a Gauss model to describe the load, in which it is assumed that the load at each hour follows a normal distribution with known mean and variance. In addition, some research shows how to transfer the load to a normal distribution [31][32][33][34], and the daily load can be divided into several segments to use normal distributions with different mean and variance. Douglas [34] thinks that it is a reasonable assumption to consider the load demand follows normal distributions because the system load is actually an aggregation of all end users connected to the grid.…”

Section: Power Purchases In a Unified Marketmentioning

confidence: 99%

“…In addition, some research shows how to transfer the load to a normal distribution [31][32][33][34], and the daily load can be divided into several segments to use normal distributions with different mean and variance. Douglas [34] thinks that it is a reasonable assumption to consider the load demand follows normal distributions because the system load is actually an aggregation of all end users connected to the grid. If each customer is considered a random variable, then their summed loads will approximate a normal distribution according to the central limit theorem.…”

Section: Power Purchases In a Unified Marketmentioning

confidence: 99%

See 1 more Smart Citation

Optimal Dispatch Strategy of a Virtual Power Plant Containing Battery Switch Stations in a Unified Electricity Market

et al. 2015

View full text Add to dashboard Cite

A virtual power plant takes advantage of interactive communication and energy management systems to optimize and coordinate the dispatch of distributed generation, interruptible loads, energy storage systems and battery switch stations, so as to integrate them as an entity to exchange energy with the power market. This paper studies the optimal dispatch strategy of a virtual power plant, based on a unified electricity market combining day-ahead trading with real-time trading. The operation models of interruptible loads, energy storage systems and battery switch stations are specifically described in the paper. The virtual power plant applies an optimal dispatch strategy to earn the maximal expected profit under some fluctuating parameters, including market price, retail price and load demand. The presented model is a nonlinear mixed-integer programming with inter-temporal constraints and is solved by the fruit fly algorithm.

show abstract

Section: Power Purchases In a Unified Marketmentioning

confidence: 99%

Section: Power Purchases In a Unified Marketmentioning

confidence: 99%

Optimal Dispatch Strategy of a Virtual Power Plant Containing Battery Switch Stations in a Unified Electricity Market

et al. 2015

View full text Add to dashboard Cite

show abstract

“…In the scenario analysis method, the uncertainty about the future demand is modeled by a number of deterministic subproblems (20) . In the other related paper, this method is also called decision tree analysis (19) , which concludes that load uncertainty imposes additional risk on short-term planning, thereby increases Expected Cost of Uncertainty (ECOU). However all the mentioned papers did not take into account the cost of uncertainty resulting from generating unit unavailability.…”

Section: Introductionmentioning

confidence: 99%

“…Many approaches in the previous works dealt with the techniques to solve the impact of load uncertainty on the UC problem based on scenario analysis (18) (19) . In the scenario analysis method, the uncertainty about the future demand is modeled by a number of deterministic subproblems (20) .…”

Section: Introductionmentioning

confidence: 99%

Short-term Operating Strategy with Consideration of Load Forecast and Generating Unit Uncertainty

Sarjiya

Eua‐arporn

Yokoyama³

2007

IEEJ Trans. PE

View full text Add to dashboard Cite

One of the common problems faced by many electric utilities concernes with the uncertainty from both load forecast error and generating unit unavailability. This uncertainty might lead to uneconomic operation if it is not managed properly in the planning stage. This paper explicitly demonstrates how to include the uncertainties to obtain the best operating strategy for any power systems. The uncertainty of the load forecast is handled using decision analysis method, meanwhile the uncertainty of the generating unit is approached by inclusion of risk cost to the total cost. In addition, three spinning reserve strategies based on deterministic criteria are incorporated in the development of scenario. Meanwhile, Mixed Integer Linear Programming method is utilized to generate unit commitment decision in each created scenario. The best strategy which gives the minimum total cost is selected among the developed scenarios. The flowchart of the proposed method is shown in Fig. 1.The proposed method has been implemented to solve a modified IEEE 24-bus system comprising 26 generating units. Three spinning reserve strategies based on deterministic criteria are set for each hour at 8%, 10%, and 12% of the load demand, representing low, medium and high spinning reserve strategy respectively. Meanwhile, the demand uncertainty is modeled by low, medium, and high Fig. 1. Flowchart of the proposed method load level with its associated probability. The load forecast uncertainty is represented by the standard deviation (SD), i.e. x% of the expected values. Sensitivity analysis with respect to the number of unit, expected unserved energy (EUE) price, standard deviation of load forecast, and probability of load level is reported.In the sensitivity analysis of number of unit, the IEEE test system is duplicated to be 26-unit, 78-unit and 130-unit system representing small, medium, and large-scale system. The load forecast error represented by SD is assumed to be 1% of the forecast value in the first hour and linearly increased to 4% in the last hour. The EUE is priced at 2000 $/MWh. The results shows that the generation cost for each scenario is a litlle bit different whereas the risk cost is rather different from each other.The impact of the load forecast error is analyzed through the variation of standard deviation (SD) values. In this analysis, it is supposed that the SD is assumed to be 1% of the forecasted value in the first hour, and is treated as either a constant value for the whole lead time of 24 hours or linearly increased to 2-4% in the last hour. The results show that with higher forecast error, i.e. higher SD, the total cost obtaied from the best strategy tends to decrease. However, as the system becomes larger, lower reserve strategy tends to be more appropiated, since the risk cost contribution to the total cost is relatively small. EUE price is varied to verify its impact on the operating strategy. It can be noticed that if the risk cost is neglected from our consideration, i.e. EUE price is set at zero, the lower spin...

show abstract

“…"OQm s}UkD u} wv w |DvU xDUO wO x@ = Q xO=iDU= R}v u} wv |=yVwQ QO w """w Rv}mvH Tm =@ 'QDr}i utr=m 'uw}UQoQ '|v=tR j]vt 'u=@}DWB Q=OQ@ u}W=t '|@Ya |xm@W Q}_v |awvYt Vwy |=ysD}Qwor= R= u=wD|t = Q """w u=oJQwt 'l}DvS 'u=oOvQB Q}_v |rt=mD |=ysD}Qwor= x= Qtyx@ |R=i "OQ@ s=v =yu; "Ov=xOQm xO=iDU= Q=@ |v}@V}B |= Q@ xQ}eDtOvJ uw}UQoQ R= u= Q=mty w TQ 'xDWPo Ca=U wO w p=L u=tR QO =tO 'xDWPo Ca=U wO w p=L u=tR QO Q=@ Q=Okt R= pOt QO xDWPo Ca=U wO w p=L u=tR QO Q@= 'xDWPo Ca=U wO w p=L u=tR QO O=@ |v}@V}B |= Q@ |v=tR |QU R= Vv=Q=mty w nv=O [5] "OvOQm xO=iDU= OwN uw}UQoQ |=ys=vx@ |v=tR |QU pOt wO w xOQm xO=iDU= QwB=ovU QO COtx=Dwm |m} QDmr=Q=@ Qy |v}@V}B |=]N "Ov=xO=O O=yvW}B |rYi ARIMA pOt w xv=oOvJ |x}RHD |QU pOt xm O=O u=Wv xrY=L G}=Dv w OW xU}=kt w x@U=Lt |v=tR |QU R= pOt wO OQmrta |= Q@ |m} QDmr= Q=@ |=[=kD COtOvr@ w COtu=}t 'COtx=Dwm |v}@V}B x@ \w@Qt |=yC}r=ai |= Q@ =y|v}@V}B "OwQ|t Q=m x@ CQOk |=ysDU}U K}LY |D}vt= |=yp}rLD w |Q=Pox}=tQU '|Q=Oxov 'CNwU O} QN 'O}rwD |=y|R}Qxt=vQ@ uDi=} |= Q@ COtOvr@`]=kt QO jQ@ |=[=kD |v}@V}B [1] "OQ}o|t Q= Qk xO=iDU= OQwt |= Q@ COtx=Dwm |v}@V}B u}vJty w QwWm |=yx=owQ}v |xaUwD Qw_vtx@ |D; |=yR=}v "CU= |QwQ[ hQYtsm w hQYtQB hrDNt |=yVN@ QO jQ@ hQYt u= R}t s}_vD |=ysDU}U |D}vt= w |O=YDk= C=}rta QO |O}rm |Wkv COtx=Dwm |v}@V}B [3] %x@ OwW|t s}UkD pwtat Qw]x@ CQOk sDU}U QO Q=@ |v}@V}B [2] [35] "OW=@ Q=PoQ}F -=D R= |@}mQD 'Cw=iDt |}=wy w ? ; w |v=mt \}=QW x@ xHwD =@ Cw=iDt C=k}kLD QO OQwt |v}@V}B xm =Hv; R= "CU= xOW xO=iDU= |v}@V}B Qw_vtx@ Cw=iDt |=yQDt= Q=B xv=oOvJ |x}RHD pOt xm |r=L QO Ovm |v}@V}B = Q =[=kD CkO x@ Ov=wD|t |v=tR [6] "CU= xOQm pta |rYi ARIMA R= QDy@ |tm QO Ow@y@ Qw_vtx@ |awvYt Vwy l}vmD R= u}kkLt R= |Q=}U@ 'Q}N= |xyO QO 1 KBES Vv=O Q@ |vD@t OvtWwy |=ysDU}U "OvOQm xO=iDU= Q=@ |v}@V}B CkO |R=i G=DvDU= sDU}U R= |}=yOQ@ Q=m , = Q}N= "OvQwyWt |=yxvwtv R= |@Ya |=yxm@W w l} [8] u= Q=mty w lQ=B [7] "CU= xDiQo Q= Qk xHwD OQwt xv}tR u}= QO |R=i |x}Q_v w xv=RwQ |m} QDmr= Q=@ |v}@V}B |xrUt |= Q@ Q=WDv=TB |x}qxU |@Ya |xm@W w u} QDu}}=B '\UwDt XN=W xU pt=W |@Ya |xm@W |OwQw "Ov=xO=O O=yvW}B |m} QDmr= Q=@ Q=Okt u} QDq=@ R}v u; |HwQN w CU= xOw@ |m} QDmr= Q=@ Q=Okt u} QDq=@ pta |v=tR |QU w |vw}UQoQ |=ypOt R= QDy@ Q=}U@ w= |O=yvW}B pOt "CU= xOw@ Q=@ |v}@V}B |= Q@ RBF x}=B |a=aW |@Ya |xm@W R= [9] l=wv "CU= xOQm w QD`} QU Q@=Q@ 11 pk=OL RBF xm CU= xO=O u=Wv w xOQm xO=iDU= |m} QDmr= R= [10] =tqU= w |r@ Q=O "CU= xOw@ Q=WDv=TB |@Ya |xm@W R= QDu=v}t]= p@=k xO=O u=Wv w xOQm xO=iDU= Czech QO |m} QDmr= Q=@ |v}@V}B |= Q@ |@Ya |xm@W "CU= xOQm pta QDy@ ARIMA pOt R= |@Ya |xm@W |O=yvW}B pOt xm CU= xOv}; Ca=U l} Q=@ |v}@V}B xm CU= xO=O O=yvW}B |@Ya |}xm@W [11] p;pO@a w xDiQo Q_v QO = Q |Da=U Q=@ w |Da=U |=tO w= &CU= xO=O s=Hv= p=U GvB |= Q@ = Q |xm@W pOt [12] [20] w=m [19] "OvyO|t u=Wv OwN R= |QDy@ OQmrta |m} QDmr= Q=@ |v}@V}B QO [21] u= Q=mty w uJ "CU= xOQm xO=iDU= |v=tR |QU xvwoJ xm OvOQm u=}@ |RH CQwYx@ =yu; "Ov=xOw@ s=oW}B SVM \UwD COtu=}t R= R}v [22] …”

mentioning

confidence: 99%