研究目的
To develop an OPF based bi-level approach for VVO to achieve CVR benefits for a three-phase unbalanced radial distribution system by simultaneously controlling both legacy devices and smart inverters.
研究成果
The proposed bi-level VVO approach effectively coordinates legacy voltage control devices and smart inverters for CVR in three-phase unbalanced distribution systems. It reduces feeder power demand by maintaining voltages near the lower ANSI limit, with higher energy savings at minimum load. The method is scalable and validated on multiple test feeders, showing computational efficiency and accuracy compared to OpenDSS.
研究不足
The approach assumes radial topology for distribution feeders, which may not apply to meshed networks. The nonlinear power flow model in Level-2 can be computationally intensive for large systems, requiring network reduction techniques. The approximations for phase angles and voltages may introduce errors in highly unbalanced conditions.
1:Experimental Design and Method Selection:
A bi-level VVO framework is proposed. Level-1 uses a MILP formulation with linear three-phase AC power flow approximation to optimize control of legacy devices and smart inverters. Level-2 uses a NLP formulation with nonlinear three-phase power flow approximation to adjust smart inverter parameters for feasibility.
2:Sample Selection and Data Sources:
IEEE 13-bus, IEEE 123-bus, and PNNL 329-bus three-phase test feeders are used for validation. Load and generation profiles are based on example profiles in OpenDSS, simulated in 15-min intervals.
3:List of Experimental Equipment and Materials:
Voltage regulators, capacitor banks, smart inverters connected to distributed generators (DGs), and voltage-dependent loads are modeled. Specific equipment includes 32-step voltage regulators, capacitor banks with binary control, and DGs with smart inverters of various capacities.
4:Experimental Procedures and Operational Workflow:
For each test feeder, Level-1 solves MILP to obtain setpoints for legacy devices and smart inverters. Level-2 solves NLP to refine smart inverter controls. Solutions are validated against OpenDSS power flow simulations.
5:Data Analysis Methods:
MATLAB is used for simulations, with CPLEX 12.7 for MILP and fmincon for NLP. Results are compared to OpenDSS for accuracy in power flow, voltages, and energy savings.
独家科研数据包,助您复现前沿成果,加速创新突破
获取完整内容