Research
For people in the power systems field, my research is best described as ‘‘data-driven distribution network optimization. The tools I have developed include parameter, topology, and state estimation techniques, as well as feedback optimization and distributed phasor feedback control, which adjust power output/load in real time in response to distribution network conditions. In addition to distribution network automation, I am also interested in electricity markets, transmission network applications, and inverter control for low inertia grids.’’
For people in the learning, control, and optimization fields, my research is best described as ‘‘online, data-driven control and optimization methods for stochastic network systems with a focus on power system applications, with additional interests in hybrid systems, market design, unsupervised learning, reinforcement learning, and graph theory.’’
For general audiences, my research is best described as ‘‘developing mathematical techniques which use new technologies such as satellite-synchronized grid sensors to help with our transition from fossil fuels to intermittent renewable energy resources by keeping track of what's happening on the grid and coordinating grid-connected batteries, electric vehicle charging, and electric home heaters/AC units.’’
Below is a list of some of the projects that I have worked on over the course of my Ph.D. and a short statement on my long term research interests.
Voltage Phasor Control
 |
Voltage Phasor Control (VPC), introduced by our team at Berkeley and made possible by high-precision Phasor Measurement Units (PMU), is a new way of controlling distributed energy resources in which the power flow optimization broadcasts voltage phasor targets to distributed feedback controllers, rather than open-loop power setpoints. We demonstrated PBC on unbalanced, three-phase distribution networks in Hardware-In-the-Loop simulations at Lawrence Berkeley Lab (LBL)'s FLEXLAB test site using the Distributed Extensible Grid Control (DEGC) software platform [DEGC paper] developed by our lab [VPC with DEGC paper] and adaptive MIMO feedback control [P-V Sensitivity presentation, MIMO Phasor Control presentation]. We have also proved that VPC is more effective than both open-loop power commands and voltage magnitude control in avoiding upstream line flow constraint violations, providing a clear use case for VPC and guidelines for how it should be applied. |
Relevant publications:
K. Moffat and A. Von Meier. Using voltage phasor control to avoid distribution network constraint violations. In 2022 IEEE Power Systems Computation Conference (PSCC), Submitted, pages 1–6. IEEE, 2022. Pre-print Link.
K. Moffat and A. von Meier. Linear quadratic phasor control of unbalanced distribution networks. In 2021 IEEE Madrid PowerTech, pages 1–6. IEEE, 2021. Paper link, PowerTech 2021 presentation.
K. Moffat Local power-voltage sensitivity and thevenin impedance estimation from phasor measurements. In 2021 IEEE Madrid PowerTech, pages 1–6. IEEE, 2021. Paper link, PowerTech2021 presentation.
K. Moffat, J. Pakshong, L. Chu, G. Fierro, Baudette J. Gehbauer C. Swartz, J., and A. von Meier. Phasor based control with the distributed, extensible grid control platform. In 2021 IEEE Power & Energy Society Innovative Smart Grid Technologies Conference (ISGT), pages 1–5. IEEE, 2021. Paper link.
G. Fierro, K. Moffat, J. Pakshong, and A. von Meier. An extensible software and communication platform for distributed energy resource management. In 2020 IEEE International Conference on Communications, Control, and Computing Technologies for Smart Grids (SmartGridComm), pages 1–6. IEEE, 2020. Paper link.
M. Baudette, J. Swartz, K. Moffat, J. Pakshong, C. Gehbauer, and A. von Meier. Hardware-in-the-loop benchmarking setup for phasor based control validation. 2021. Paper link.
Grid Parameter Estimation
 |
The lack of accurate distribution network models and knowledge of the operating state prevent most DER dispatch methods from being deployed in real life. We have proved that, while it is impossible to estimate the full network admittance matrix from voltage and current phasor measurements, it is always possible to estimate the admittance matrix for the Kron-reduced network [Limitations and Tools paper]. We also demonstrated that estimating the ‘‘subKron’’ impedance matrix, motivated by Laplacian graph theory, is more robust to measurement noise because the data is better conditioned. We have also implemented the least squares complex data regression recursively, eliminating the need for storing and processing large batches of PMU data [Real Time paper]. Subsequent work, in collaboration with researchers at EPFL/ETHZ, has leveraged a Bayesian Error-In-Variables formulation to produce the first accurate, reproducible 3-phase distribution grid estimation from PMU measurements with realistic signal-to-noise ratios without active perturbation [Bayesian Error-In-Variables preprint]. |
Relevant publications:
K. Moffat, M. Bariya, and A. Von Meier. Unsupervised impedance and topology estimation of distribution networks—limitations and tools. IEEE Transactions on Smart Grid, 11(1):846–856, 2019. Paper link.
K. Moffat, M. Bariya, and A. von Meier, Network Impedance Estimation For Microgrid Control Using Noisy Synchrophasor Measurements, 2018 IEEE 19th Workshop on Control and Modeling for Power Electronics (COMPEL), 2018, pp. 1-6 Paper link.
K. Moffat, M. Bariya, and A. Von Meier. Real time effective impedance estimation for power system state estimation. In 2020 IEEE Power & Energy Society Innovative Smart Grid Technologies Conference (ISGT), pages 1–5. IEEE, 2020. Paper link.
J.-S. Brouillon, E. Fabbiani, P. Nahata, K. Moffat, F. Dorfler, and G. Ferrari-Trecate. Bayesian error-in-variables models for the identification of distribution grids. IEEE Transactions on Power Systems, In Revision, 2022. Preprint link.
State and Topology Estimation
 |
Electric grids have switches which reconfigure the network topology. Considering an alternative grid operation scenario in which a network model is available but the switch statuses are unknown, we introduced the Multiple model Adaptive Power system State Estimator (MAPSE) [MAPSE presentation], which models the distribution network as a dynamic hybrid system — the voltage phasors are the continuous state variables, while each discrete ‘‘mode’’ corresponds to a different network configuration. The MAPSE estimates the hybrid state by 1) maintaining a continuous state estimate for each discrete mode using a bank of Unscented Kalman Filters (UKFs), and 2) estimating the topology/discrete mode by applying Bayesian recursion on the UKF residuals. We demonstrated that the MAPSE can accurately detect network topology changes using PMU with noise levels consistent with real PMU data [Empirical paper]. |
Relevant publications:
K. Moffat and C. Tomlin. The multiple model adaptive power system state estimator. In 2021 IEEE Conference on Decision and Control (CDC), pages 1–6. IEEE, 2021. CDC 2021 presentation.
M. Bariya, K. Moffat, and A. Von Meier. Empirical noise estimation in distribution synchrophasor measurements. In 2019 International Conference on Smart Grid Synchronized Measurements and Analytics, pages 1–7. IEEE, 2019. Paper link.
Feedback Optimization for EV Charging Control
 |
I have been working with teams at the National Renewable Energy Lab (NREL) and (LBL) on the Department of Energy's GEMINI-XFC project to coordinate EV charging on oversubscribed distribution networks. The demonstration is a full-detail transportation network-and-distribution grid co-simulation of the Bay Area. Thus, the control algorithm must work at scale. I am implementing Feedback Optimization algorithm, which runs the optimization algorithm in closed loop with the physical system. Computational effort is saved because the Feedback Optimization uses the physical grid to ‘‘solve’’ power flow. |
Longer Term Perspective
Engineers, computer scientists, and mathematicians are grappling with how to use massive data, computing power, and communication to improve automated decisions for network systems such as electric grids. Data-driven control methods such as (deep) reinforcement learning have demonstrated impressive results in simulation, but applying data-driven control to safety-critical systems remains an open problem. For safety-critical systems, the guarantees that come with principled mathematical approaches are paramount.
I am interested in developing theory and tools that bridge the gap between simulation and real-world implementation. In particular, I am interested in:
combining online optimization with distributed feedback control,
incorporating a priori model knowledge and intuition into unsupervised learning methods, and
integrating unsupervised learning with online control/optimization.