Research On Unit Commitment For Power Systems With High Penetration Of Wind Power

This dissertation focuses on two dimensions of research on the unit commitment for the power systems mainly constituted by thermal generators and integrated with large-scale wind power,including the optimization approach and thermal generator's flexibility,so as to solve the two kinds of problems caused by large-scale wind power integration,including increasing complexity of unit commitment and enlarging flexibility requirements.The dimension of optimization approach studies the complexity of unit commitment caused by the modeling of wind power characteristics;based on the research on the optimization approach,the dimension of thermal generator's flexibility expands the models of the thermal generators and evaluates the impacts of the flexible operation of thermal generators on the unit commitment.This dissertation includes the following four parts:(1)This dissertation proposes the reserve models considering the stochastic characteristics of wind power,which can concurrently ensure the economic efficiency and avoid the huge computational burden,and thus effectively solves the problem faced by the current reserve optimization research on wind power.The newly proposed reserve models convert the complex correlation between both the costs and benefits of reserve capacity into a more concise and easy-to-model physical association and thus decouple the tight relationship between the complexity of reserve optimization and the number of stochastic scenarios.Therefore,the reserve models can be constructed through a large number of wind power scenarios to accurately reflect the impact of the stochastic characteristics of wind power on reserve optimization,which is beneficial to the unit commitment to accurately evaluate the costs and benefits of reserve capacity and thus to ensure the economy of the reserve strategy.Meanwhile,this dissertation has verified that the reserve models have at least an approximately convex characteristic and thus is convenient for piecewise linearization using the piecewise linear functions of the piecewise accumulation form.After the piecewise linearization,these complex and non-linear reserve models can be transformed to a small set of linear constraints constituted by continuous variables,so as to greatly reduce the computational burden.(2)This dissertation proposes a new extended deterministic network-constrained unit commitment approach,which solves the reserve optimization problem and the network congestion problem faced by the unit commitment with wind power through the proposed reserve models and the proposed simplified network constraints,respectively.This approach effectively solves the difficulty faced by the current optimization approaches in concurrently achieving both economic performance and computational efficiency objectives.This approach incorporates the reserve models into the deterministic unit commitment framework to concurrently optimize the system reserve levels and on/off decisions,so as to comprehensively evaluate the costs and benefits of reserve capacity as the typical stochastic approach does,and thus has the economic performance similar to the typical stochastic approach.Meanwhile,this approach preserves the advantage of the deterministic approach in the computational performance because these reserve models just bring in very little computational burden.In addition,the proposed approach compensates the disadvantage of the typical deterministic approach in network congestion analysis through the proposed simplified network constraints and avoids the huge computational burden caused by a large number of inactive network constraints adopted in the typical stochastic approach.(3)This dissertation proposes an accurate thermal generator model considering the startup and shutdown power trajectories and studies the impact of the accurate modeling of the startup and shutdown processes of thermal generators on the unit commitment with wind power regarding both the economy of solution and computational efficiency.This study has verified the economic value of the accurate modeling of the startup and shutdown processes of thermal generators and proposed the solution to overcoming the possible heavy computational burden caused by accurate modeling.The newly proposed thermal generator model just takes into account the critical phases of the startup and shutdown processes so that this accurate model can not only reflect the startup and shutdown trajectories of thermal generators but also avoid the introduction of a heavy calculation burden.In addition,this study reveals that the accurate modeling of startup and shutdown processes has a different impact on the unit commitment.Accurate modeling of the startup process causes a heavier computational burden but has a weaker contribution to economic benefits.Therefore,it is advisable to simplify the accurate modeling of the startup process when the unit commitment problems face the heavy computational burden.(4)This dissertation proposes the fuel cost model considering low-load operation with auxiliary firing(LLOAF)to express the operating costs of coal-fired generators adopting LLOAF mode,proposes the economic evaluation method of LLOAF,and studies the impact of the LLOAF of coal-fired generators on the unit commitment with wind power regarding both the economy of solution and computational efficiency.This study has made up for the lack of the current theory in the universal modeling and economic evaluation method of LLOAF,verified the economic value of LLOAF of coal-fired generators,and proposed the solution to overcoming the possible heavy computational burden caused by the modeling of LLOAF mode of coal-fired generators.This dissertation transforms the complex and piecewise non-linear fuel cost model considering LLOAF into a set of mixed-integer linear constraints,thus greatly reducing the computational burden.Combining with the proposed fuel cost model,this dissertation proposes the economic criterion and economic index to judge whether it is economic for a coal-fired generator to be operated in the LLOAF mode and thus provides a theoretical reference for assessing the economy of LLOAF.This study reveals that LLOAF can improve the economic performance of the solution of unit commitment and the proposed extended deterministic approach can effectively reduce the increasing computational burden caused by the fuel cost model considering LLOAF.