Optimization method for reducing network loss of dc distribution system with distributed resource

With mature of the direct current (dc) distribution grid technology, the distributed energy is embraced as much as possible, and reducing network loss becomes an effective way to increase energy efficiency and system stability. The conventional reducing loss approaches include complex algorithms, loss modeling in power devices and adding hardware, which lead to costs and system response time increase and system stability reduce. Considering the Intermittent characteristics of distributed energy, for reducing network loss, the effective option is to control error and response speed directly through the upper control, when distributed energy accesses on dc distribution network, which is no need to change grid topologies, run cumbersome power flow algorithms, nor add additional equipment. A reducing network loss method is proposed based on this idea. The details are as follows: Firstly, the network loss formula is derived based on power flow calculation, and the network loss change rules are analyzed. Secondly, one optimal power flow mathematical model of dc distribution network is established, which takes the minimum network loss as the objective functions and conforms to constraints of system security and components operating limit. The tide optimization is solved by using the artificial bees (ABC) algorithm. Thirdly, the network loss reduction method is proposed in the dc distribution network by using master–slave control through real-time control instruction optimization. The node voltage, branch current and main voltage source converter power are precisely regulated to control the power flow; thus, the network loss of multiple distributed energy access dc distribution network can be reduced. Fourthly, when wind and solar energy is connected to the dc distribution network, a typical IEEE16 node case is demonstrated to verify the feasibility of the proposed method using software MATLAB/SIMULINK. Fifthly, evaluation and prospect are made for the research.