Study Of Unit Commitment In Power System Based On Robust Optimization Theroy

Unit Commitment(UC)problem is mainly applied to decide the optimal schedule of the power generation for various units in power system,so that make it meet the load demand at minimum cost(or maximum profit)and various constraints of security,resources,environment,etc.Efficient UC Schedule is the foundation for secure,economical and environmentally friendly operation of power system.For its importance and complexity,UC problem has been an important topic which aroused extensive concerns in both engineering and academic fields for several decades.A number of new factors,such as large quantities of renewable power energy represented by grid-tied wind power,N-K contingencies,power market etc.,bring more uncertainties to UC problem which itself actually is already a very difficult problem,and make it to be more complicated than it ever before.Meanwhile,the social gradually growing dependence on electrical power supply and the open loop operation mode for depressing the over high short-circuit current due to tighter connection of power system,make the contradiction between power supply reliability and depressing the short-circuit current more prominent.How to configure a UC schedule scientifically with proper consideration of the fore-mentioned uncertainties and balance the conflict between short circuit current and power supply reliability level is a research subject with not only great significance in theory,but also urgent demand in practical engineering field.To focus on the above issues,from the aspects of basic theory study and practical engineering application,an overall and in-depth research for UC problem based on robust optimization theory has been done in this thesis,including basic robust optimization theory improvement,rational modeling with multi-uncertainties and efficient solving method.A series of expected excellent results and achievement have been gained in this thesis consist of following.Firstly,a novel Improved Light Robust Optimization(ILRO)model and its solution algorithm based on sorting-and-truncation method are presented in the thesis.The model proposed has better objective value by letting limited constraint violations;the improvement in conservatism of ILRO is given in the form of proposition and proved strictly.For the robust optimization model with uncertainties on right hand side and the uncertain parameters are subjected to a budget uncertain set,a sorting-andtruncation method is designed to deduce the linear counterpart of robust constraint directly,which lead the ILRO to be solved more efficiently;and the equivalence between linear counterpart of robust constraint and its original one is strictly proved by proposition.The above achievement contributes the development of robust optimization theory.Secondly,a novel and less conservative Wind-Thermal UC model based on ILRO is proposed and its corresponding Mixed Integer Linear Programming(MILP)counterpart,which is more tractable,is derived.In the model,light robust constraints are designed by considering wind power dispatching and allowing electrical equipment to be short term overload,which significantly reduce the solution conservatism to make the operation cost be lower.The primal MILP counterpart of the robust constrains for the model derived by sorting-and-truncation method can be solved more effectively by linear solution solvers,compare with traditional dual transferring method which has to adopt nonlinear solution solvers due to the appearance of the nonlinear terms.Thirdly,a novel co-optimization model of short circuit current constrained unit commitment problem and transmission switching(TS)problem is built.The short circuit current of the system can be easier to be limited,since the union optimization with both measures of generators up-down and transmission switching are adopted.Meanwhile,the optimal network operation mode can be gained as well.The short circuit current constraints in UC problem are firstly modeled by utilizing node admittance matrix,and the linearization method for them is further presented,which keep the model linear and could be solved with linear method.The results of the samples demonstrate the validity of the model and algorithm proposed.Fourth,a comprehensive robust optimization model for UC-TS with both the wind power and N-K contingency uncertainties and short circuit current constraints,and its algorithm are further proposed.Because of a large amount of decision variables,uncertain parameters,and constraints,the model is very complicated and is difficult to solve directly for large scale system.Two decomposition algorithms with some simplifying strategies are presented in this dissertation.Algorithm 1 is the two-stage solving method based on the combination application of Big M method,outer approximation and benders decomposition,which is relative easy to be implemented but only suitable for small scale system.Algorithm 2 is the two-stage and multi-level decomposition solving method,in which the original model is initially simplified by fixing the daily switching operation strategy and filter out the inactive short circuit current constraints,then the model is decomposed by C&CG method into two-stage problems: The first stage problem is a certain UC-TS problem with short circuit current constraints,which is further decomposed by benders decomposition into three relative smaller scale problems,in which more effective benders cuts and anti-islanding cut sets are also generated and added later;the second stage problem is a robust optimization based re-dispatching problem with uncertainties,outer approximation method is utilized to process the appearing nonlinear items due to dual transformation.The two stage problems are combined through C&CG cuts and iteratively solved until the convergence is achieved.Computation results of the IEEE 118 system demonstrate that the algorithm above is valid,and its calculation precision and speed meets the requirement of the practical engineering.Therefore,the algorithm above owns the potential to be applied for practical engineering.