The Bridge - Issue 1, 2022 - 24

Graduate Research SP
TLIGHT
Samuel Talkington
Beta Rho, Georgia Institute of Technology, Ph.D. Student in Electrical and Computer Engineering
Ph.D. Candidate in Aerospace Engineering, NASA Space Technology Research Fellow
RESEARCH TOPIC
Physics-Based, Data-Driven Modeling of Solar PV and Smart Inverters
Utilities rely on software models of electric power systems for planning
interconnections of new energy sources. These models are typically
derived from the physical laws governing power system behavior. Today,
private " behind-the-meter " installations of solar photovoltaics (PV)
are expanding at a rapid pace, creating unpredictable changes in grid
topology and increasing the risk that a system model may be inaccurate.
Samuel's work, supported by Sandia National Laboratories, develops
detection and estimation methods that meet the challenge of unreliable
system models in an energy landscape with high levels of PV integration,
helping to expedite society's transition to a renewable energy portfolio.
Specifically, his most recent work develops new techniques to estimate
the power factor control setting of a PV system's inverter, which controls
the ratio of active and reactive power supplied. In contrast with traditional
approaches, the method reconstructs this setting without directly
measuring the power injections. By integrating physical power system
laws with high-dimensional statistical estimation methods, the inverter's
active and reactive power injections are reconstructed and separated
using just voltage magnitude measurements from the grid, which
are more readily accessible. These methods reduce the need
for future deployment of expensive energy measurement devices on
the electric grid, and can improve model accuracy in rural or
underserved communities.
LEARN MORE
https://talkington.dev/
Figure 1: The real power voltage
sensitivity matrix for a power system
model. This matrix, based on power
system physics, describes how the
voltage measurements (M) respond
to active power injections at different
locations (L) in the system.
Figure 2: The reactive power voltage
sensitivity matrix for a power system
model. This matrix has subtle differences
in comparison to the active power
voltage sensitivity matrix in Fig. 1. Using
them together allows for the power
factor to be reconstructed.
Figure 3: A visualization
of reconstructing the
power factor control
setting of a distributed
generator (e.g. Solar
PV) using the voltage
sensitivity matrices. (95%
confidence interval bars
shown). Injections of
real (left) and reactive
(right) power can be
independently separated.
CONTACT
https://www.linkedin.com/in/samtalkington/
Jacopo Sini
Mu Nu - Politecnico di Torino, Politecnico di Torino, Italy
Ph.D. Candidate in Control and Computer Engineering, M.S. in Mechatronic Engineering, B.S. in Computer Engineering
RESEARCH TOPIC
Novel Validation Techniques for Autonomous Vehicles
The development of highly dependable microprocessor-based embedded systems is a challenging activity. These systems
are fundamental to making vehicles more flexible and efficient, less polluting and safer. Unfortunately, these systems raise
functional safety issues since they have to react on time and appropriately, even if hardware failures affect their components.
It is required that these systems be reviewed by two analyses: Hazard Analysis and Risk Assessment (HARA) to determine
the risk levels associated with the hazards involved in the requested functionalities and Failure Mode, Effects, and Diagnostic
Analysis (FMEDA) to investigate how the possible failure modes of their components affect functional safety.
Jacopo's work examined how to implement novel approaches to perform the HARA and
FMEDA analyses to improve their repeatability and objectivity. He proposes methods to aid
these analyses thanks to the adoption of simulation systems and fault models. In particular,
in FMEDA, it is possible to assess the effectiveness of the fault detection, isolation, and
recovery (FDIR) mechanisms implemented, resorting to embedded software. Real-time
software testing techniques are used to validate the implementation on the target platforms.
Teaching activities aimed at university students and company employees is a fundamental
part of his work to disseminate these techniques and safety culture.
LEARN MORE
https://scholar.google.com/citations?user=5gopV8QAAAAJ
CONTACT
jacopo.sini@polito.it
A screenshot from CoppeliaSim
during the FMEDA of a mobile robot.
Real-time software
testing harness
THE BRIDGE

https://www.talkington.dev/ https://www.linkedin.com/in/samtalkington/ https://scholar.google.com/citations?user=5gopV8QAAAAJ

The Bridge - Issue 1, 2022

Table of Contents for the Digital Edition of The Bridge - Issue 1, 2022