Search

As modern electric power distribution systems (MEPDS) continue to grow in complexity, largely due to the ever-increasing penetration of Distributed Energy Resources (DERs), particularly solar photovoltaics (PVs) at the distribution level, there is a need to facilitate advanced operational and management tasks in the system driven by this complexity, especially in systems with high renewable penetration dependent on complex weather phenomena. Digital twins (DTs), or virtual replicas of the system and its assets, enhanced with AI paradigms can add enormous value to tasks performed by regulators, distribution system operators and energy market analysts, thereby providing cognition to the system. DTs of MEPDS assets and the system can be utilized for real-time and faster-than-real-time operational and management task support, planning studies, scenario analysis, data analytics and other distribution system studies. This study leverages DT and AI to enhance DER integration in an MEPDS as well as operational and management (O&M) tasks and distribution system studies based on a system with high PV penetration. DTs have been used to both estimate and predict the behavior of an existing 1 MW plant in Clemson University by developing asset digital twins of the physical system. Solar irradiance, temperature and wind-speed variations in the area have been modeled using physical weather stations located in and around the Clemson region to develop ten virtual weather stations. Finally, DTs of the system along with virtual and physical weather stations are used to both estimate and predict, in short time intervals, the real-time behavior of potential PV plant installations over the region. Ten virtual PV plants and three hybrid PV plants are studied, for enhanced cognition in the system. These physical, hybrid and virtual PV sources enable situational awareness and situational intelligence of real-time PV production in a distribution system.

The power distribution system is disorderly in design and implementation, chaotic in operation, large in scale, and complex in every way possible. Therefore, modeling, operating, and controlling the distribution system is incredibly challenging. It is required to find solutions to the multitude of challenges facing the distribution grid to transition towards a just and sustainable energy future for our society. The key to addressing distribution system challenges lies in unlocking the full potential of the distribution grid. The work in this dissertation is focused on finding methods to operate the distribution system in a reliable, cost-effective, and just manner. In this PhD dissertation, a new data-driven distributed ($D^3M$) framework using cellular computational networks has been developed to model power distribution systems. Its performance is validated on an IEEE test case. The results indicate a significant enhancement in accuracy and performance compared to the state-of-the-art centralized modeling approach. This dissertation also presents a new distributed and data-driven optimization method for volt-var control in power distribution systems. The framework is validated for voltage control on an IEEE test feeder. The results indicate that the system has improved performance compared to the state-of-the-art approach. The PhD dissertation also presents a design for a real-time power distribution system testbed. A new data-in-the-loop (DIL) simulation method has been developed and integrated into the testbed. The DIL method has been used to enhance the quality of the real-time simulations. The assets combined with the testbed include data, control, and hardware-in-the-loop infrastructure. The testbed is used to validate the performance of a distribution system with significant penetration of distributed energy resources.

State estimation is a fundamental task in power system monitoring. The focus in this thesis is on Distribution System State Estimation (DSSE). One of the main challenges in DSSE is the lack of observability due to the small number of sensor installations in practical power distribution circuits, where the number of measurements is far fewer than the number of state variables. In this thesis, our goal is to develop DSSE methods which address the low-observability challenges.First, we leverage the high reporting rate of a small number of distribution-level phasor measurement units (D-PMUs), a.k.a., micro-PMUs, to unmask and characterize sparsity patterns among the state variables in radial power distribution systems. Accordingly, the DSSE problem is formulated over the differential synchrophasors as an adaptive group sparse recovery problem to track the changes that are made in the states of the system and captured by D-PMU measurements. To enhance the performance of the proposed method, the formulated DSSE is further augmented by the side information on the support of the vector of unknowns that is obtained from the outcome of an event-zone identification analysis prior to solving the DSSE problem. Second, to capture the dynamics of the power distribution system, we model the DSSE problem under an event-triggered setting, where we use the estimations of the state variables during the previous events as priori information to predict the state variables at the current event. Accordingly, a novel data-driven method based on elastic net regression is proposed to learn the event-triggered state transition matrix; despite the low-observability in the system. Here, in the absence of direct power measurements, we enhance our ability in sparse recovery by developing a new reinforced physics-based coupling method among the state variables, in which we add a novel set of linear differential power flow equations to the DSSE problem formulation in forms of virtual measurements. Third, we study the joint estimation of sensitivity distribution factors and power flows in low-observable power distribution systems by developing a novel physics-aware measurement-based approach that takes into account the sparsity features of the problem extracted for radial power distribution systems.

Power distribution systems have underwent a lot of significant changes in the last two decades. Wide-scale integration of Distributed Energy Resources (DERs) have made the distribution grid more resilient to abnormal conditions and severe weather induced outages. These DERs enhance the reliability of the system to bounce back from an abnormal situation rather quickly. However, the conventional notion of a radial system with unidirectional power flow does not hold true due to the addition of these DERs. Bidirectional power flow has challenged the conventional protection schemes in place. The most notable effects on the protection schemes can be seen in the field of islanding or Loss of Mains(LOM) detection and general fault identification and isolation. Adaptive protection schemes are needed to properly resolve these issues. Although, previous works in this field have dealt with this situation, a more comprehensive approach needs to be taken considering multiple topologies for developing adaptive protection schemes. The most common protective devices widely deployed in the distribution system such as overcurrent relays, reverse power relays at Point of Common Coupling(PCC), fuses, reclosers and feeder breakers need to studied in implementing these schemes. The work presented in this dissertation deals with simulation based and analytical approaches to tackle the issues of islanding and adaptive protection schemes. First we propose a multiprinciple passive islanding detection technique which relies on local PCC measurements, thus reducing the need of additional infrastructure and still ensuring limited Non Detection Zone (NDZ). The next step to islanding detection would be to sustain a islanded distribution system in order to reduce the restoration time and still supply power to critical loads. Such an approach to maintain generator load balance upon islanding detection is studied next by appropriate shedding of non-critical and low priority critical loads based upon voltage sensitivity analysis. Thereafter, adaptive protection schemes considering limited communication dependency is studied with properly assigning relay settings in directional overcurrent relays (DOCRs), which are one of the most widely deployed protective devices in distribution systems by catering to multiple topologies and contingencies. A simulation based technique is discussed first and then an analytical approach to solve the conventional optimal relay coordination problem using Mixed Integer Linear Programming (MILP) with the usage of multiple setting groups is studied. All these approaches make the distribution more robust and resilient to system faults and ensure proper fault identification and isolation, ensuring overall safety of system.

Tornadoes are hazards of low probability of occurrence and high consequences that cost the United States billions of dollars each year. Electric power distribution systems are susceptible to damage due to tornadoes with the utility poles being the most vulnerable components. Additionally, the reliability of power distribution systems can be affected by the deterioration of the strength of utility poles with age. Many utility companies nowadays are considering the use of steel and prestressed concrete poles instead of wood poles, which are the most widely used in the United States. Up to date, very few studies have been performed to study the behavior of power networks when subjected to tornadoes. This research proposes a framework to perform reliability analysis, cost analysis, and target hardening of power distribution systems subjected to tornado hazard. It also offers a framework to compare the reliability of wood, steel, and prestressed concrete utility poles subjected to tornadoes through fragility analysis considering the deterioration of the strength of the poles with age.

Recent developments in the area of power generation have led to an increased penetration of storage and distributed energy resources (DER) in power distribution systems. As a result, new and enhanced energy management systems will be necessary as the deployment of DERs, as well as the need to control loads, continues to increase in the coming years. Advanced management systems are especially important to achieve resilient power delivery during emergency situations. During a blackout for example, a section of a distribution feeder could island to operate as a microgrid to ensure critical services such as water, food, and medical care remain online. The development of such management systems will also require the ability to integrate human behavior models with power flow simulators, as technological advances lead to more customer-owned devices with the potential to be used to balance the power flow of a feeder. This thesis describes a co-simulation framework that combines a bottom-up residential load generator, a load aggregator for real-time (RT) residential demand response (DR), a utility-scale battery controller, and the GridLAB-D distribution system simulator. The behavior of a distribution feeder model is analyzed under different scenarios. The model is based on an existing feeder located in Los Alamos, NM, which serves residential customers and a set of critical loads including a hospital, a supermarket, and a water treatment plant. This feeder also hosts a utility-scale solar array and battery storage that are used to operate the feeder as a microgrid. Additionally, a real-time simulation is described in which real-time residential demand response is implemented on a virtual community using the load generator and aggregator. The simulation is part of a project invested in the development of a modern microgrid control system employing a virtual power plant approach and a model predictive controller to optimize the use of resources within a distribution feeder. The capabilities to study power distribution systems with humans in the loop using these platforms is showcased here. Furthermore, their potential as instrumental tools in the development and design of new technology essential to improve grid resilience is also discussed.

Power distribution systems have been significantly affected by a wide range of faultcausing events; and the current outage restoration procedure may take from tens of minutes to hours. Effective outage cause identification can help to expedite the outage restoration and consequently improve the system reliability. Most current researches are based on system modeling and measurements such as voltage and current; besides, they usually target at a single feeder or a small system due to the difficulty of modeling the large-scale, nonlinear, and time-varying distribution system. In this research, various data mining approaches including statistical methods and artificial intelligence algorithms have been investigated and applied to Duke Energy distribution outage data in order to extract the outage pattern and identify the outage cause; by this means, the additional environmental information recorded in the data can be adopted in the fault diagnosis and the analysis range can be beyond the scope of a single feeder or a small system. Also, the affect of data imperfections such as data noise, data insufficiency, especially the data imbalance issue on the performance of outage cause identification have been investigated. In this work, logistic regression and artificial neural network are firstly compared on their capability in fault diagnosis; then an existing fuzzy classification algorithm is extended to Ealgorithm to alleviate the effect of data imbalance; afterwards, the immune system based Artificial Immune Recognition System (AIRS) algorithm is investigated for its capability in fault diagnosis using real-world data; lastly, a hybrid algorithm based on E-algorithm and AIRS is proposed to embed the rule extraction capability while performing satisfactory fault cause identification.