IEEE Power & Energy Magazine - January/February 2022 - 36

SCE has also expanded the deployment of more novel
technologies such as distribution fault anticipation and early
fault detection. These technologies leverage a combination of
electromagnetic and acoustic signal patterns generated from
potential system anomalies to preemptively identify leading
indicators of potential system issues. Once these issues are
identified, SCE can proactively remediate the issue before
an actual failure occurs.
An example of an incipient condition identified by
these technologies is broken strands on overhead conductors.
These types of defects are essentially nondetectable
with conventional protective relays and will usually
remain in the field, with the condition continuing
to deteriorate until an actual failure occurs. With early
fault detection, SCE can pick up the anomaly from the
acoustic noise generated from the broken strands (Figure 9).
A pair of neighboring sensors pickup similar signals and
based on the timestamp of each sensor on both sides of
the conductor, SCE can narrow in on the location of the
incipient fault and dispatch field crews to confirm the
potential problem and remediate it before it deteriorates
to the point of failure.
Additionally, artificial intelligence and machine learning
are being applied to develop more sophisticated algorithms
to help identify equipment defects. The machine
learning algorithm is examining a large volume of photo
images acquired through SCE's aerial inspection process
using helicopters and drones in potential high firerisk
areas. With thousands of images captured daily, it
could take weeks before problems are identified. With the
advances in the machine learning algorithm, the examination
process for certain assets is automated, thereby
identifying potential defects very quickly (Figure 10).
SCE continues to invest in building additional machine
learning models to detect as many different asset defects
as possible.
Risk Modeling and
Investment Priorities
One of the challenges of executing wildfire mitigation
measures is determining the right suite of measures that
should be implemented and what locations should be prioritized
for deployment. With ~27% of SCE's large service
area identified as exposed to elevated or extreme
wildfire risk, SCE needs appropriately prioritize the
locational deployment of system-hardening measures to
reduce the greatest amount of wildfire risk in the shortest
amount of time.
Risk Models Enable Better
Decision Making
In 2020, SCE met some significant milestones in enhancing
risk analytics, including:
✔ integrating an enterprise-level risk modeling approach
with asset- and location-specific risk models
✔ transitioning to a new ignition consequence modeling
tool that uses more up-to-date and expanded historical
data at a higher granularity
✔ developing asset-specific probability of ignition models
for transmission and subtransmission assets in addition
to the distribution asset models built previously
figure 9. An example of an incipient system anomaly
identified by using early fault detection technology.
✔ supplementing its wildfire risk model to include
PSPS as part of the overall risk, thus more accurately
accounting for risks impacting customers
and risk reduction associated with wildfire mitigation
activities.
These improvements enable SCE to drive consistent riskDEFECTIVE_CROSSARM
99%
informed decision making at the enterprise and activity
levels, help more accurately estimate risk along the grid
and to communities, and better target how much work to do
where and when.
Risk Analysis Allows an Efficient
Allocation of Resources
SCE has performed risk-reduction and risk-spend efficiency
figure 10. An example of a machine learning algorithm to
quickly identify a cross-arm defect from pictures acquired
through aerial inspections.
36
ieee power & energy magazine
calculations using the granular approach mentioned in the
" Risk Assessment and Mapping " section. This provides
a more accurate understanding of relative risk buy-down
with any WMP activity and enables SCE to evaluate the
relative risk-reduction benefits of WMP activities more consistently.
SCE is using the results of risk analyses to make
january/february 2022
IEEE Power & Energy Magazine - January/February 2022

Table of Contents for the Digital Edition of IEEE Power & Energy Magazine - January/February 2022
