Allocation of Dump Load in Islanded Microgrid Using the Mixed-Integer Distributed Ant Colony Optimization

Abstract

In highly penetrated microgrids, the problem of voltage and frequency deviations exceeding their permissible limits, becomes significant with higher share of renewable based distributed generation. The existing real-time control systems of highly penetrated microgrids cannot cope on its own with such deviations, owing to the large generation and demand mismatch mainly during off-peak hours. A dump load can help with voltage and frequency regulation by consuming excess generation. However, further investigation is required to highlight the importance of optimal dump load allocation on the operation of microgrids. The mixed-integer distributed ant colony optimization is introduced as a novel application in droop controlled islanded microgrids to minimize voltage and frequency deviations and system losses. The optimization problem was formulated as a single-and multi-objective problem to allocate a dump load and the optimal droop settings for distributed generation in islanded microgrid during off-peak hours. The proposed optimization method was teamed up with a special backward/forward sweep load flow method to account for distributed generation droop characteristics and enhance the solution convergence. The method was applied to the IEEE 69-and 118-test systems and validated against competitive swarm and evolutionary metaheuristics. Results have shown that an optimally sized and allocated dump load with optimized droop setting could minimize voltage and frequency deviations to an acceptable level, while reducing power losses incurred by the installation of such load into the microgrid.

Dump Load Allocation in Islanded Microgrid with Robust Backward\Forward Sweep and MIDACO

“…In DCIMG, generation units are typically modelled as inverter-based DG (IBDG), where 𝑉𝑉 -𝑓𝑓 droop control is facilitated by power electronics. The inverter control system of typical IBDG utilizes the 𝑃𝑃-𝑓𝑓 and 𝑄𝑄-𝑉𝑉 droop equations given in (1) and (2), respectively [4]. Such equations influence DGs output active (𝑃𝑃 𝐺𝐺𝐺𝐺 ) and reactive (𝑄𝑄 𝐺𝐺𝐺𝐺 ) power to enable load change following as per IEEE std.1547.7 [13]:…”

“…) Conversely, for sake of simplicity, a constant power load model was adopted for all loads and DL in this study [4]. To account for droop response by DGs the SBFS method was proposed in [4]. SBFS consists of four stages as follows:…”

“…Dump loads (DL) are known to provide adequate power management solution for significant mismatch situations [2]. Nonetheless, few studies have tackled the optimal DL allocation problem to provide 𝑉𝑉-𝑓𝑓 support at low load hours in DCIMG framework [2]- [4]. Therefore, more studies are necessary to highlight the importance of DL planning in IMG.…”

“…However, ill-conditioned systems with variations in droop setting or generation and load levels degrade the convergence of the forgoing methods and limit the boundary conditions for decision variables of optimization problems. As in [4], stability improvement in highly penetrated MG was achieved by the optimal dump load (DL) allocation using combination of robust LF method called special BFS (SBFS) and the mixed-integer distributed ant colony optimization (MIDACO) algorithm. The use of stochastic optimization techniques to solve non-convex mixed-integer non-linear problems (MINLP) has gained significant interest in many years [1].…”

“…This is done by finding the optimal DL size and location as well as the optimal droop setting for DGs to minimize 𝑉𝑉-𝑓𝑓 deviations and power losses in the IEEE 33-bus islanded system. The first LF method, called improved SBFS (SBFS-II), is based on SBFS [4] with enhanced convergence rate. While the second method, called general BFS (GBFS), employs two dynamic damping factors with a correction vector to offer accurate representation of droop-control with enhanced convergence.…”