Parameter Tuning for LV Centralized and Distributed Voltage Control with High PV Production

Abstract

This paper presents a centralized and a distributed voltage control strategies for a low voltage (LV) grid-connected microgrid with high penetration of photovoltaic (PV) systems. Both strategies are formulated as an optimal power flow (OPF) problem to minimize line losses and PV curtailments. Second-Order Conic Programming (SOCP) relaxation is formulated to "convexify" the OPF problem. Alternating direction method of multipliers (ADMM) is implemented for the distributed voltage control strategy. Moreover, weighting parameters for both controllers are introduced to have more flexibility in the controller operating mode. A sensitivity analysis on those weighting parameters allows to identify different operating regions, with priority given to loss reduction or curtailment minimization. In particular, we identify an infeasible region in cases where the considered SOCP relaxation is not valid, which highlights the need for fine parameter tuning of the controllers. The effectiveness of both voltage control strategies is validated in a LV radial microgrid.Index Terms-alternating direction method of multiplier (ADMM), optimal power flow (OPF), photovoltaic (PV) system, second order cone relaxation, voltage control

A Three-Stage Strategy With Settlement for an Energy Community Management Under Grid Constraints

“…However, if any bus voltage violation is detected, a second control step with voltage control is formulized as a multiobjective optimal power flow (OPF) problem with secondorder conic programming (SOCP) relaxation, adopted from previous works [31]. Moreover, it is formulized as a distributed problem, with the local/global variables of household h corresponding to the actual RT power dispatch (p netRT h,t…”

“…The OPF aims to minimize the global forecast mismatch and grid losses, with θ l , θ n , ∈ [0, 1] being the controller parameters that weight the two objectives. Note that, these weighting parameters have to be tuned properly to ensure the feasibility and the desired operating point of the algorithm [31]. The SOCP relaxation convexifies the equality relation of the current (23b) by transforming it into inequality relation [25].…”

“…On the other hand, the CM will reward the households and pay them for their energy exported to the community. This CM's cash flow can be modelled as community budget balance (31).…”

“…The community surplus in (31) can be represented and utilized by the CM in different ways. For instance, the CM can set the surplus for its personal revenue or investments, by adjusting the difference between the buying and selling prices.…”

“…Additionally, (32b) is proposed to limit household's total cost and ensure it will not be higher than a conventional case where households trade their energy fully with the retailers. In the proposed allocation problem (32), the CM can freely set the community surplus in (31) or the penalty deviation price (π dev ) as well as the grid service price (π gs ) can either be optimization parameters or variables. This topic will be discussed further in the next section.…”