IEEE Power & Energy Magazine - May/June 2019 - 45

accurate estimate of the state is not required. instead, a sufficiently accurate approximation to detect the cyberphysical mismatch early enough in the execution of the dc power
order is necessary.
the preceding discussion is concerned primarily with
cyberthreats originating in the Scada network at the converter station level. however, at the wide-area level, detection
and mitigation are becoming increasingly more important
for preventing systems from being affected by cyberattacks
targeting measurements and control commands transmitted
through WaMpac networks. this is particularly important
for those WaMpac applications requiring real-time data
processing and control.
Because WaMpac networks are geographically dispersed over thousands of kilometers, complicated methods
that rely on system-wide data analysis can hardly satisfy
stringent timing requirements, due to the extensive execution time for data processing and the time delay introduced
by networking and time alignment. Similarly, mitigating
cyberthreats at the system level cannot always guarantee
that the disturbances induced by cyberattacks are mitigated
before they result in severe consequences. Furthermore,
the timing performance is critical for security functions to
detect incorrect commands that could drive the system to
unstable conditions. in this sense, a detection and mitigation strategy implemented at the pdc level is more suitable
to protect the hvdc-based WaMpac application in real
time. the pdc controlling an hvdc station can provide
the input for the hvdc controller through a lan, so the

pdc only needs to subscribe to signals from remote pMUs
and pdcs. the detection and mitigation functions can be
implemented on the pdc and integrated with the time
alignment and correction functions, thus minimizing both
execution and data transmission time.
an example detection and mitigation framework for
hvdc-based WaMpac applications is shown in Figure 5. the pdc controlling the hvdc station receives
and time-aligns the pMU data stream, then sends the
data to the detector and controller. the detector can be
operated either as a built-in function in the pdc or as
a separate device in parallel with the pdc. the detection and mitigation algorithms inside the detector rely
only on time-aligned pMU data streams from the pdc
and can be independent of other pdc functions. the output of the detector includes the detection status for each
time-stamped data sample. if the data sample is found
to be compromised, the detector will calculate an estimated signal to replace the compromised signal. here, the
input selector functions as a data source switch between
the pdc and the controller, switching to the estimated
data generated by the detector only under attack conditions. Under normal conditions, the input selector will not
introduce any delay. Under attack conditions, because the
detector also provides estimated signals, a small amount
of delay will be introduced by the detector to calculate
estimated signals. the hvdc-based WaMpac controller
should be able to respond to the detection status and adapt
to the estimated signals.

dc
Set Point Value
Command

Offline Analysis
* Find the average number of
iterations for different values
of dc power set point for fast
state estimation.

Power System 1
Control Center
SCADA /EMS
dc
Measurements

All
Good

Yes

3
Measured Power ==
Estimated Power?

ac Line-Flow
Measurements for All
ac Lines

Fast State
Estimation to
Determine Power
Injection on dc Bus

* PTDFs will be determined
based on the power transfer
between the sending and
receiving end on the HVdc
system. PTDFs will be used to
determine the transmission
lines that need to be used for
the fast state estimation
approach.

No
Malicious Command

figure 4. The detection and mitigation of WAMPAC-originated cyberthreats. PTDF: power transfer distribution factor.
may/june 2019

ieee power & energy magazine

IEEE Power & Energy Magazine - May/June 2019

Table of Contents for the Digital Edition of IEEE Power & Energy Magazine - May/June 2019