Studies On The Key Technologies Of Active Distribution Network Planning

Due to the increasing concerns on energy crisis worldwide,the power grid,one of the major energy consumption networks,is developing towards a renewable,clean,and sustainable system.Designed for electricity distribution and consumption,the power distribution network is also changing from “fit and forget” mode to active and controllable operation pattern,resulting in active distribution systems.Active distribution network is treated as the most promising form of future distribution systems and shows complicated features resulting from the renewable energy integrations,active managements and diverse energies inter-connections.Therefore,there are new challenges for power grid planning.This dissertation focuses on the studies of active distribution network planning based on the key characteristics including uncertainty,operation controllability and energy inter-connection.The research framework based on bi-level optimization is proposed to take all characteristics into consideration,and the vital technologies in power grid planning are concluded and formulated within the framework,while the corresponding researches are conducted.Specifically speaking,following contents are involved:Considering the stochastic features caused by the wide integrations of distributed renewable energies,researches are conducted from two views: non-sequential analysis for statistical features based on probability density and sequential analysis for dynamic features based on scenario technologies.As for non-sequential aspect,the adaptivebandwidth kernel density estimation approach is proposed at the first place,to handle the abnormality of probability density curves of renewable power injections.Secondly,probabilistic load flow based on randomized quasi-Monte Carlo sampling and Johnson transformation is presented,leading to the fast and accurate analysis of uncertain power system and providing an efficient solution to stochastic planning.As for the sequential aspect,the uniform design based dynamic scenario analysis approach is proposed,and the representative scenarios characterizing the stochastic features of variables are formulated by generating the samples in a uniform way.And furthermore,the impacts of different control strategies based on diverse equipment,such as electric vehicle or energy storage,are introduced by scenario modification or expansion approaches,providing an effective method to introduce active management into generated scenarios.Considering the controllable features introduced with the active management system,an “upper-planning and lower-operation” based bi-level optimization model is proposed to properly characterize and integrate the active management in the network planning.In the upper level,the main concern is the grid planning,while a stochastic programming model based on slacked chance constraints is presented to handle the random features of renewable energies.Furthermore,the system uncertain analysis is addressed by the randomized quasi-Monte Carlo sampling and Johnson transformation based probabilistic load flow.In the lower level,the main concern is the network operation.The optimal control for typical scenarios generated by the uniform design based dynamic scenario analysis approach is investigated.In order to quantify the bilevel optimal value,the interval number satisfaction criterion is introduced in the particle swarm optimization algorithm to evaluate the objectives and provide the optimal solutions.As a result,uncertainties and sequential control patterns are properly addressed in active distribution network planning by aforementioned studies.Considering the energy inter-connections due to the energy efficiency concerns,the active distribution network planning with energy inter-connections is formulated,where the gas-turbine based energy internet is taken as an example.In the beginning,the models of nature gas systems and combined heat and power generation systems are formulated based on the principles of power system analysis,and the unified energy flow analysis approach is proposed following the power flow analysis scheme,which is an effective method dealing with energy internet issues.Next,state excursion criterion and virtual power to heat ratio are proposed and corresponding control strategy for interconnected energy systems is also presented,which are all used in the modelling of active management and energy system planning.Finally,the multi-objective planning model is introduced to handle the complicated features of hybrid systems,while all objectives of subsystems are set based on their own characteristics and the bilevel optimization model with upper level for planning and lower level for operation is deployed for optimization.As a result,a scientific planning approach for energy internet is formulated.