Distributed Stochastic Economic Dispatch For Integrated Transmission And Distribution Networks With Renewable Energy

In recent years,with the upgrade of the approaches and concepts of Chinese economic development,the role of renewable energy in the energy structure of China is increasingly notable,and it has become one of the primal goals for the operation and dispatch of power system to accommodate the output of renewable energy in priority.Since the output of renewable energy is strongly volatile and stochastic,storage devices such as pumped-storage hydro stations and battery storage stations are employed for power balance adjustment.Meanwhile,a considerable number of distributed power sources has been integrated into the distribution network for the past few years.How to fulfill the power generation potential of these sources and realize better accommodation of renewable energy on the level of integrated transmission and distribution networks(ITDN)has become a key concern for the dispatch decision-making agencies.Therefore,it is necessary to study the distributed stochastic economic dispatch problem of the ITDN with renewable energy.First of all,study on the security-constrained unit commitment(SCUC)problem of transmission network with 110 k V lines is conducted.Based on the evolution of the structure of most urban power grids in China,it is proposed to rearrange the division point of ITDN at the110 k V bus of the 110/10 k V substation.Furthermore,since the X/R ratio of 110 k V lines are quite small,direct power flow model may incur huge errors in computing the active power of lines,therefore,a novel quadratic active power flow model are introduced and convexified by the means of convex hull theory.Results on the test cases demonstrate that the proposed model can obtain solution as precise as that of the alternating current power flow model with less computational time consumption.Secondly,study on the distributed stochastic economic dispatch problem of the ITDN with renewable energy is conducted.In order to reduce the difficulty of solving the dispatch problem of distribution network,linear power flow(LPF)model is proposed based on the Taylor series expansion,and the state constraints of storage devices with 0-1 discrete variables are continuous linearized.On this basis,transmission network with wind farms and PSH stations and distribution network with solar plants and BS stations are modelled separately,and the stochastics of renewable energy’s output are considered based on approximate dynamic programming(ADP)and the storage model theory.Successive projective approximation routine is applied to train the approximate value functions with error scenarios.In order to solve the dispatch problem of ITDN distributionally,synchronous alternating direction method of multipliers are combined with the ADP algorithm to formulate the distributed stochastic algorithm.Results on the test cases verify that the proposed LPF model and the utilization of continuous linearization can evidently reduce the time consumption to solve the problem.Additionally,the proposed distributed algorithm can achieve a higher computational efficiency than the traditional scenario method and centralized algorithm by temporal and spatial decoupling and obtain a more satisfactory dispatch results that meet the practical requirements.Finally,in view of the potential losses caused by the extreme scenarios of uncertain fluctuation of renewable energy’s output,the risk-averse distributed stochastic algorithm is put forward.Based on the distortion risk measures(Glue Va R)that can simultaneously consider multiple scenarios,a cost term that reflect the impact of extreme scenarios to the objective function of the problem.Coherent risk measure theory is applied to simplify the mathematical expectation term,and the training process of approximate value functions is modified.Results on the test case indicate that the proposed algorithm can obtain corresponding solution according to different preference of risk of the dispatch decision-making agencies,achieving a balance between the economy of the forecast scenario and the conservativeness of extreme scenario.