| Under the background of new-type power system,the penetration of renewable energy sources(RESs)with strong uncertainty is increasing,which affects the safe and economic operation of power system while realizing clean and low-carbon.Effectively handling the uncertainty caused by RESs urgently awaits to be solved.In the most of optimal scheduling research in active distribution network(ADN),the uncertainty sets are pre-specified,i.e.,the uncertain information is often used as a parameter.However,the increasing uncertainty challenges the current optimal scheduling models.It is necessary to soften rigid uncertain information,that is,treat uncertainty as an object to be optimized,quantify the optimal accommodated uncertainties in ADN from the perspective of the regulation capacity of ADN,reasonably build the disturbance resistance capability of ADN,and provid a dispatch for RESs to reduce the system operation risk and improve the system operation economy.Although there have been some studies on the admissible region of RESs and load in transmission system,it has not been studied in ADN,and the admissible region optimization models for transmission system cannot be simply and directly duplicate to ADN.Therefore,with the optimal scheduling of ADN as the theme,this paper studies how much uncertainties can be handled in ADN under the existing infrastructures.The main work and achievements are as follows:(1)Three linearized power flow applicable to distribution network are introduced firstly.Based on these three linearized power flows,three new ISFs applicable to distribution network are derived respectively,so that the active/reactive line flow and nodal voltage can be directly expressed quantitatively by nodal active/reactive power injections.And the new ISFs can help solve the following proposed models.(2)This paper introduces the net load admissible region,and proposes an economic dispatch model for ADN considering net load admissible region.In this model,the indrect uncertainties of net load are considered.To ensure that the reserve can be fully exploited and successfully delivered after the realization of uncertainties,the indirect uncertainties of line flow and nodal voltage caused by net load are also considered.The generator base points,participation factors,reserve and nodal admissible region can be optimized simultaneously.By using the preemptive goal programming,the model with two objection functions is transformed into a two-level model,so as to maximize the admissible region and improve the economy as much as possible.The new ISF and traditional affine policy are used to avoid the solution impediment caused by uncertainties.The affinely adjustable robust optimization model is transformed into a deterministic nonlinear model by using dual theory.Finally,the validity and effectiveness of the proposed model are illustrated on a modified IEEE 33-node distribution network and a modified IEEE 69-node distribution network.(3)This paper proposes a dynamic economic dispatch model for ADN considering conditional value-at-risk(CVaR)of net load.Based on the cost/benefit analysis,risk cost is introduced into the objection function to strike a balance between operation cost and operation risk.A surrogate affine policy is used to avoid the introduction of bilinear terms and improve the computation efficiency.Finally,the affinely adjustable robust optimization model is recast as a deterministic mixed integer linear problem,and the validity and effectiveness of the proposed model are illustrated on a modified IEEE 3 3-node distribution network and a modified IEEE 69-node distribution network.(4)This paper proposes an economic dispatch model for ADN based on a feasible region projection method.Unlike the two models described above,which use affine policy and ISFs to eliminate random quantities,the basic idea of this model is to eliminate the random variables using Fourier-Motzkin Elimination.Finally,the model is transformed into a deterministic linear programming model and the validity and effectiveness of the proposed model are illustrated on a modified IEEE 33-node distribution network. |
