Konrad Schmitt

After initial development and testing with synthetic data, distribution systems solutions are required to be verified under diverse network topologies and consumption patterns to confirm their deployment feasibility. Even though the literature has presented data for several test distribution networks over the years and smart metering datasets have been made available, finding them according to the research requirements is time-consuming. Especially in reconfiguration and restoration studies, test networks must have a networked topology, which makes the search for models even more limited. Still, the available network models and datasets tend to be oversimplified or omit crucial data and users are often left with the need to make assumptions. This article compiles and presents eight different reconfigurable distribution networks ranging from 12 to 328 nodes and nine smart metering datasets for residential. Along with the compiled data, this article discusses the available and missing information of each model and dataset, while potential assumptions and their impacts are discussed. With that, the main intention of this article is to provide a complete and centralized resource to support the development of various active distribution networks’ studies.

Advanced distribution management systems (ADMSs) embrace several online and offline monitoring and control features to ensure power delivery with reliability, continuity, and quality. By having data about the system’s topology, ADMS can obtain real-time field measurements to perform further analysis on the network and automatically respond to events or support operators’ decisions. Academia has been advancing the current state of the art of ADMS by proposing more sophisticated solutions that aim to address the ongoing changes in active distribution networks. However, such solutions are commonly tested in offline environments, neglecting the dynamics and challenges related to real-world field equipment and communication protocols. The most faithful power systems solution can get tested before actual deployment is through hardware-in-the-loop (HIL) testing, where electrical systems are simulated in real-time digital simulators, and physical equipment is integrated into the simulator through analog hardwiring and communication protocols. Considering the need for more realistic testing and validation of proposed solutions to facilitate real-world deployment, this research proposes and presents an HIL testbed for ADMS’s testing purposes that may guide future studies in the area. Along with the testbed, an automatic outage restoration ADMS feature is formulated and tested on the proposed HIL system, where the benefits of pushing validation to the most faithful testbed are highlighted.

Maintaining service reliability, achieving sustainability, and ensuring energy efficiency are crucial for green high-performance computing systems. Balancing these factors is a key challenge for modern green data centers. In this research, we propose a monitoring, visualization, and management system for green data centers (MVMS-GDC). Our comprehensive automated platform includes “monitoring system” and “rules and policy management” modules. The “monitoring system” gathers and visualizes time series data from all resources of a green data center, tracking essential metrics and measurements. It audits green energy, microgrid, climate conditions, workloads, hardware, CPU and memory usage, cluster component health, computing node activities, and network health and quality metrics. The “rules and policy management” module defines and enforces policies to balance resources, ensuring a reliable, sustainable, scalable, and efficient computing environment. We implemented, tested, and evaluated the MVMS-GDC system using green energy at the Zephyr data center located at the GLEAMM site. Our results demonstrate at least a 4.9% improvement in performance, at least a 4% increase in energy efficiency, and a reduction of at least 4% in job losses. The MVMS-GDC system also enhances scalability by employing a policy machine for each compute node, which automates power state control (on, off, or hibernation) based on monitoring observations. This automated approach ensures efficient and dynamic scaling, making MVMS-GDC suitable for large and highly distributed data centers. Overall, MVMS-GDC provides a robust solution for balancing energy availability and computational needs, optimizing performance, and maintaining energy efficiency in green data centers.

Efficient reduction of power losses in modern systems can be achieved by selectively connecting tie lines and disconnecting radial branches in distribution networks with scattered energy generation sources. The influence of load power is pivotal in this optimization process. Consequently, it becomes crucial to accurately model the demand side, portraying the behavior of electricity consumers more effectively. Given that loads are not constant and exhibit variations with changes in voltage magnitude, contingent on the consumer type, it becomes imperative to represent each load through its load model's constant-impedance, constant-power, and constant-current components. Establishing the relationship between these components and load types is essential for proposing more flexible reconfiguration models. Therefore, this study employs mathematical formulations to express the correlation of load components with consumer types, aiming to provide a more flexible reconfiguration formulation compared to conventional models. The results obtained from well-known distribution systems demonstrate that the proposed framework is sufficiently efficient, offering optimal solutions within a shorter computing time compared to other existing reconfiguration models.

Today, distribution networks are important parts of modern power systems that include specific numbers of distributed generation (DG) units. Although DG utilization in distribution grids diminishes energy dissipation, integration of DG into these networks is constrained by financial and technical restrictions. To more effectively reduce these losses, distribution systems should be reconfigured to adapt to their variable and unpredictable load demands and renewable energy generation. Nonetheless, these uncertain amounts affect energy losses notably. Thus, the proposed configurations should not be modified by every change in load and generation levels (i.e., the reconfiguration approach must be relatively robust against load and generation uncertainties). Also, taking demand variability into account when reconfiguring distribution systems enhances the complexity, computational load, and time of processing. Therefore, strategies of reconfiguration should be selected for their robustness and quick convergence in the face of load and generation variability and uncertainty. Hence, this study developed an effective and robust strategy for reconfiguring distribution grids considering variable and uncertain load and renewable generation. Some of the most important features of the presented approach are its simple implementation, appropriate robustness, and relatively short computational time, as shown in simulation results.

Distribution networks often suffer from substantial energy losses, particularly in radial feeders. These losses, primarily in the form of wasted heat, far exceed those experienced in the transmission system. Such power dissipation not only drives up operating costs but also degrades the quality and quantity of delivered electricity. One effective strategy to curtail these losses is the optimal placement of shunt capacitors. By compensating for a portion of the reactive power demanded by loads, these capacitors reduce the magnitude of current flowing through feeders and subsequently diminish distribution losses. However, the electricity demand within these networks fluctuates over time, significantly influencing the power distributed through the feeders. These load variations play a crucial role in determining energy losses, reactive power injections, and system's total cost. The dynamic nature of load variability can necessitate adjustments to the optimal locations for capacitor placement. However, taking into account the entire load variations can substantially increases the complexity and computational time required for solving the capacitor placement problem, particularly in large distribution systems. In light of these challenges, current study introduces a highly effective formulation for optimal capacitor placement to minimize energy losses and capacitor installation costs in distribution systems with time-varying loads. The numerical analysis of the results underscores the effectiveness of our proposed framework in achieving optimal capacitor bank places in radial distribution feeders operating under variable load conditions.

The reconfiguration of distribution networks results in modifications to the voltage magnitudes of buses, subsequently influencing the power demands of connected loads. This variable load demand plays a pivotal role in determining the optimal system topology. While a few studies have recognized this crucial aspect, they often present non-linear models or simplified formulations. Non-linear models, though solvable with commercial solvers, demand extensive computational time, particularly when accommodating the time variability of loads in the formulation. Alternatively, solving these models may require meta-heuristic methods, introducing the challenge of obtaining precise solutions. In contrast, existing linear models have been subject to approximation through piecewise linearization, introducing a level of uncertainty. Achieving accurate solutions for reconfiguration problems within a short computational timeframe is imperative for practical operational applications. As a result, present study addresses this challenge by introducing four effective reconfiguration methods that explicitly consider the voltage dependence of time-varying loads. These techniques leverage linear solvers, enabling the computation of precise solutions within a condensed computational timeframe. By adopting these proposed methods, the study aims to enhance the efficiency and accuracy of solving distribution network reconfiguration problems, thereby facilitating practical applications in real-time operational scenarios. The linear mathematical models introduced in the current study can find exact solutions in a short computational time, unlike meta-heuristic algorithms and non-linear formulations. Additionally, the proposed time and voltage variability formulation causes the presented reconfiguration frameworks to solve the problem in very low computational time compared to models with exact load variations inclusion.

Enhancing the efficiency of distribution systems can be effectively achieved through hybrid reconfiguration and the coordinated operation of distributed generators (DGs). This integrated approach proves to be highly effective in minimizing power losses within distribution networks, with a particular emphasis on the substantial impact of load power and grid voltage on distribution losses and the output of DGs. Load and DG powers are dynamic, varying with fluctuations in voltage levels. Furthermore, the correlation between power demand and renewable sources generation with voltage is contingent upon the specific type of load and DG. Nevertheless, scant attention has been given to these crucial factors in the research pertaining to network reconfiguration and the planning of DG. A limited number of papers have taken into account the type and voltage dependency of DG or load within their respective formulations. However, the reconfiguration models they proposed introduce significant nonlinearity without considering the simultaneous effects of grid voltage on load and DG power, necessitating computation through nonlinear solvers or metaheuristic algorithms. Alternatively, the implementation may require intensive linearization techniques for compatibility with linear solvers. Nonlinear solvers demand extensive computational time, while metaheuristic algorithms cannot ensure the attainment of optimal solutions. Hence, it is crucial to accurately model load behavior when undertaking the reconfiguration of active distribution systems. This paper presents a streamlined reconfiguration model, designed to be easily implemented with conventional optimization tools while ensuring accuracy in identifying appropriate solutions for the reconfiguration problem.

The operational cost of distribution systems is significantly affected by the energy losses of the system. Distribution network losses are known to be higher than transmission system losses, constituting over 70% of all the losses that occur in the power system. For the achievement of loss minimization, operation cost mitigation, and generation expense reduction, the efficient allocation of distributed generation (DG) in distribution networks is considered an effective solution. The penetration of renewable DG units in distribution systems has seen an increasing trend in recent times due to the numerous benefits that come with the use of renewable sources for power generation, especially for the environment. Power demand variations and the variability of load levels play a vital role in the determination of the amount of losses that occur on the network and subsequently the optimal generation of DGs. Meanwhile, taking into consideration the variability of load level in the DG allocation problem worsens the computational stress and time significantly. Therefore, this paper has formulated an effective DG allocation model for energy loss minimization considering power demand variation. The developed model was simulated for different distribution systems using A Mathematical Programming Language (AMPL) to validate the performance of the proposed formulation by comparing the proposed model results with two different models (exact and peak load model). Parameters tested in the study are network losses, cost, and computational time. The evaluation of numerical results indicates the proposed model is very effective for DG allocation in terms of losses, cost, and computational time.

Distribution networks are susceptible to de-energizing many customers during an outage event. Through distribution network reconfiguration (DNR) techniques, outage events can be partially or entirely restored by rerouting power to customers in a new configuration that also isolates the faulted string. To improve reliability indices, utilities have been deploying many reclosers per feeder, increasing the chances for protection miscoordination. Based on that, it is crucial that the restoration solution not only finds an optimal switching configuration to re-energize customers but also identifies miscoordinations in the system to narrow the faulted region to the smallest string possible. This paper presents a two-stage distribution systems outage restoration solution. Its novelty integrates a set of logic to properly localize the faulted area while identifying potential miscoordinations between reclosers to then improve the optimal outage restoration benefit. The restoration is computed as a Multi-Objective Optimization (MOO) based on a Mixed-Integer Convex Programming (MICP) model that minimizes unserved customers and switching maneuvers while maintaining de-energized the faulted area, providing the appropriated switching control sequence (SCS) to achieve the optimal set of open and closed switches. The validation is made in a realistic distribution systems model where different outage locations and miscoordination conditions are considered.