Research On Power Flow Optimization Of AC/DC System With DC Distribution Network

With the increasing of DC loads and the need for integrating multiple Distributed Generations economically,the traditional AC distribution network operation mode is faced with great challenges.The DC distribution network can directly provide DC grid-connected interfaces for electric vehicle charging stations and solar power stations,reducing redundant power conversion links,thereby reducing energy loss during power conversion and reducing redundant investment in power conversion equipment.If all DC loads are powered by their respective rectifiers,it will inevitably lead to huge equipment investment and huge waste of energy.However,if the DC distribution network is added to the original AC network,which forms an AC/DC hybrid system with DC distribution network,in this case,the AC/DC hybrid system can not only provide the grid connection interface for the AC/DC power source and the AC/DC loads at the same time,but also can reduce the energy loss and lower the pressure of environmental protection.However,the DC distribution network with distributed generation and energy storage units,as an active network,will change the power distribution of the original AC system when incorporated into the AC network.If the DC distribution network is directly integrated into the AC system without proper optimization,it may cause the voltage issues of the AC system.In addition,the uncertainty of grid-connected power of solar power generation units and electric vehicles in DC distribution networks makes the operation and planning of AC/DC hybrid systems more difficult.In this context,the following research is carried out on the power flow optimization of AC/DC hybrid systems with DC distribution network:Firstly,the steady-state power flow model of DC distribution network,voltage source converter and AC system is studied,and the control strategy of AC/DC system with DC distribution network is analyzed,and its steady-state power flow is established.The correctness of the established steady-state power flow model is verified by an example.Secondly,based on the steady-state power flow model of the AC/DC system with DC distribution network,the characteristics of the control variables in the system are analyzed,and the appropriate optimization algorithm is selected to optimize the static power flow of the AC/DC system with DC distribution network.Then,due to the fact that the AC/DC system with DC distribution network contains a large number of random variables,the uncertainty of the random variables in the AC/DC system with DC distribution network is taken into account by the probabilistic optimal power flow calculation based on the three-point estimation method.In this way,the beneficial search space of some control variables in the dynamic optimal power flow is determined in advance.Finally,since the DC distribution network contains energy storage units with time accumulation effects,the generators in the AC system also need to consider the climbing rate constraints.Therefore,based on the established static optimal power flow model,the equation and inequality constraints of energy storage unit are added,as well as the climbing rate inequality constraints of the generator,establishing a dynamic optimal power flow model of the AC-DC system with DC distribution network.Using the beneficial search space of some control variables determined in advance by the probabilistic optimal power flow calculations,to perform the dynamic optimal power flow calculations of the AC/DC system with DC distribution network,and the optimal scheduling instructions can be obtained.In view of the existence of weak voltage buses in the AC/DC system with DC distribution network,the voltage level of the AC/DC system is effectively improved by taking into account the V2 G mode of the electric vehicles.In addition,the probabilistic optimal power flow and the dynamic optimal power flow calculation are performed again under the droop control strategy.The results prove that the proposed method of predicting the search space of some control variables in advance through the probability optimal power flow calculation is versatile and effective for improving the dynamic optimal power flow optimization result.