Embedded Files
My research goals are to continue and expand my research work in the field of smart grid. For this research goal, I will focus on pursuing the following research areas in the near-future:
Volt-VAR Operations – Modeling of offline/online tools to enable the coordinated operation of regulating devices such as tap changers, Dynamic VAR compensators (DVCs), capacitor banks, and real power management of energy storage, demand response, smart inverters, etc. for varying conditions. Develop new control strategies for mitigation of high penetrations of PV solar by focusing on the coordination among different regulating devices. Key modeling challenges to tackle would include developing models that accurately represent the internal controls of smart inverters and constraints on their VAR supply/absorption, dynamic models of inverters and their controls. Investigate new Volt-VAR optimization techniques to solve high renewable penetration issues in distribution systems while considering power flow solutions. In my previous research, I have investigated a distributed control strategy for large scale PV penetration during cloud movements. The high ramp-rate event requires a fast and appropriate level of action from a dynamic regulating device, as conventional regulating devices fail to operate in time. High ramp rates from PVs cause the voltage to rise proportionally as shown in the fig. 1. A smart inverter was utilized for the control action by communicating with the existing slow regulating devices. In the developed distributed algorithm as shown in fig. 2, the active power control from the smart inverter was added to the reactive power support for mitigating the voltage rise and fall due to the high ramp output from the PV. To absorb and supply the imbalance of the required PV inverter output and the PV array output, a battery was utilized. A weighted mean of the current state of charge of the battery dictated the length of the controlled ramp output from the inverter. The coordinated strategy between the smart inverter and the existing regulating devices proves to be an effective solution to maintain good power quality as well as minimize the tap operations that increases its lifetime service operation. In the future, I would plan to explore intelligent multi-agent distributed control strategies to coordinate the smart inverters and regulators for wide-spread DERs of different sizes at different nodes using sensitivity factors. I have worked on a multi-stage optimal operation of cascaded regulators and smart inverter tied to PV sources to maintain voltages within a band and minimize the over usage of regulating devices. A coordination strategy between the regulating devices in the feeder was introduced. A zonal level-based management of regulating devices utilizing Volt/VAr control methods was adopted. The novelty of this work was to distribute a large multi-objective problem (due to numerous regulator and inverter control objectives) into smaller subproblems (one cycle of operation) based on a voltage regulator’s regulating zone/area. The non-linear problem was transformed into a mixed integer quadratic programming (MIQP) problem, which made it possible to attain optimal solutions in near to real-time.
Fig. (a) Reactive power capability of inverter (b) Voltage dependence on active power output at the PV bus and (c) PV active power output.
Fig. Flowchart of the proposed distributed voltage control strategy
Advanced Distribution management system (ADMS) – Testbed development by modeling the electrical network in a software environment to reduce the R&D costs before field deployment. Utilize real-time simulators/hardware testbed to perform interconnection studies while performing hardware testing to investigate system impacts. For example, a microgrid controller hardware/software model can be tested in a simulated power system via power hardware-in-the-loop (PHIL) or controller hardware-in-the-loop (CHIL). Previously, I had worked on the “Advanced Distribution Management System (ADMS) Testbed Setup” project, for which I performed data parsing from OpenDSS files to input excel files for OPAL-RT. This work enables the use of open-source feeder models to be converted to software model in a real-time simulator to test vendor-neutral technologies. In the future, I plan to investigate the validation of ADMS functionalities by using both simulation and hardware tools in a laboratory environment before they can be deployed in the field. Investigate different communication frameworks such as centralized, decentralized and distributed topologies for monitoring and control. Consider the possibility of utilizing different sensors, PMUs, telemetry devices to achieve faster communication speeds for critical operations from different nodes in the electrical network for greater observability in cases of high DER penetration. Investigate distributed communication frameworks, such as the OpenFMB[1] to achieve interoperability of devices from different vendors when being utilized for control and protection use cases.
Fig. Proposed NREL ADMS Testbed (source: https://www.nrel.gov/esif/assets/pdfs/distributedcontrols_kroposki.pdf)
Fig. Open-source model file conversion to real-time simulator interfaced files
Microgrids – Key modeling gaps include dynamic models of distributed generation controls and hardware needed to ensure reliable transition from on-grid to off-grid and vice versa. Additionally, adaptive control and protection hardware/software schemes need to be further researched for microgrid operation both in the grid-connected and island modes. Investigate the control and protection schemes for the fast response of distributed generation in microgrids and the coordination that is needed for a reliable operation. Investigate optimization techniques, energy/demand management, markets and power system planning of distributed energy resources. I have previously worked on the power electronics controls of distributed energy resources in microgrids to maintain stability for varying conditions using EMTP tools such as PSCAD and RSCAD. I worked on a project to perform microgrid operation for power regulation for an islanded operation based on the time of operation in a day. For on-peak period, a two-stage optimization strategy based on economic and frequency deviation objectives was developed. Using the microgrid energy management system (M-EMS), an economic generation scheduling was performed based on a secondary optimization control strategy with the help of several droop enabled dispatchable distributed generators (DDGs) and hybrid systems. When the batteries reached their minimum state of charge (SoC), the batteries got disconnected, allowing the inverters to operate in grid feeding mode to supply power based on the MPP of renewable distributed generators (RDGs). During the off-peak period, a multi-stage optimization strategy was suggested to charge the batteries back to their nominal values and perform regular operation. The battery utilization was leveraged based on the time of day and usage patterns, as the charging was performed during lower load periods and power supply operation was performed during the high load periods. Results showed the proposed scheduling strategy based on the microgrid system characteristics would benefit in scheduling distributed generation efficiently.
With an increasing number of natural disasters and potential cyber-attacks, both the Department of Energy (DOE) and the Department of Defense (DOD) are seeking solutions to enhance critical infrastructure resiliency. In line with the current opportunities to attract external funding, I plan to extend my research to networked microgrids for the management of several microgrids via tie-switches in an electrical distribution system with high penetration of renewables and energy storage. The different microgrid architectures which include serial microgrids on a single feeder, parallel microgrids on a single feeder, and interconnected microgrids on multiple feeders, can be formed and regulated by utilizing Energy Management Systems (EMS) and microgrid controllers. Techno-economic research studies to be performed on networked microgrids would include ancillary service management, feeder reconfiguration/restoration, protection schemes, power balancing for variable renewable energy, quantification and allocation of the benefits of networked microgrids, energy markets based on a distribution system operator (DSO), and communications architectures.
Load and PV Modeling – I have worked on a technique to extract the variability information from a distribution transformer historical data-sets, by applying frequency-based transforms. The expected similar load profiles were grouped together and decomposed using discrete wavelet transform (DWT). A classification technique was applied to cluster groups that have similar variability by observing the strong correlation coefficients among the rapid-refresh histograms. A look-up table was created to use as few classes as possible to create synthetic load profiles. The synthesized data sets were compared to the training data sets for the variability matches. Although this work dealt with residential load modeling, it is expected that the approach will be applicable to both commercial and industrial customers, as well as PV and wind profiles. In the future, I plan to utilize frequency-based transforms and machine-learning techniques to forecast power output from solar PV systems and aggregated load consumption. Explore algorithms to minimize errors associated with forecasting PV generation and load consumption using historical data, current weather parameters such as cloud movements, temperature, rain, humidity, etc. ahead of time. These forecasted profiles could them be fed to perform optimization on the electrical system under uncertainty.
Fig. Distribution feeder serving residential loads at the secondary side of the distribution transformers with high-speed DMUs.
Fig. Time-series profiles (4 hours) of the original and synthetic data set for the four seasonal months.
Google Sites
Report abuse