Event Analysis

Hands On Relay School

2019

Event Analysis

Copyright © SEL 2019

APP751_EventAnalysis_r8 1
 Event Analysis

Ahhh, the Good Old Days…

I wonder where
the problem
is…
The Good Old Days?

Photo courtesy of Rafael Garcia, ONCOR Electric Delivery

Copyright © SEL 2019

In the aftermath of the Northeast Blackout of 2003, congressional action granted the Federal Energy
Regulatory Commission (FERC) and the North American Electric Reliability Corporation (NERC)
the power to implement mandatory and enforceable reliability standards. Whether these efforts and
their costs are improving power reliability is debatable; recent misoperation statistics are level and
generally show no improvement. In 2010, NERC revitalized its event analysis program; however, the
industry lacks safe harbor guarantees that self-reported information will not be used to punish and
fine the reporters. The North American Transmission Forum (NATF) promotes the idea that the
candid exchange of information is the key to improving reliability; however, it is a members-only
group, exclusive of manufacturers, regulators, and other key stakeholders. Individual utilities work
on root-cause analysis and improving their own processes; however, sharing lessons learned with
others in an efficient way is difficult. Following outages, especially at industrial plants, the threat of
lawsuits further hinders the sharing of information.

Significant improvements in reliability will likely not occur until the industry emulates models such
as those provided by the Commercial Aviation Safety Team (CAST) and the nuclear power and
healthcare industries. Rather than blaming individuals, collaborative expert system-think teams
inclusive of all stakeholders focus on human factors, system issues, and improvements. CAST, in
particular, has amazing results: an 83 percent reduction in the fatal accident rate over 10 years as well
as improved productivity (lowered costs), all without a single new regulation.
 Event Analysis

What Is an Event Report?

• Snapshot of the power system
• Trip or trigger prompts data capture
• Includes:
– Date and time stamp

– Sampled analog inputs

– Status of inputs
and outputs

– State of relay
elements and logic

– Relay settings
Copyright © SEL 2019

Analysis of event reports and SER data can provide the root cause of protection system operations so
that you can identify if equipment can quickly be restored or if testing and repairs are required.

You will want to analyze event reports frequently and use them as a tool to improve protection and
maintain vital equipment.

The relay captures 15 or 64 cycles of event data and creates an event report, which includes four
sections:

• Analog values of current and voltage (translated into waveforms)

• Digital states of the protection and control elements and digital inputs/outputs

• Event summary (trip type, date, and time)

• Settings in service at the time of the event

APP751_EventAnalysis_r8 3
 Event Analysis

Relay Signal Path

Data Retrieved at Different Points

Copyright © SEL 2019

The relay can retrieve two types of event data: either 4 samples/cycle filtered data or
16 samples/cycle unfiltered (raw) data.

The data are retrieved at different points in relay processing, as shown above in the data path for the
voltage and current input signals.

After the currents and voltages are reduced to acceptable levels by the instrument transformers, the
signals are filtered with an analog low-pass filter, which rejects higher harmonics. Then it is digitized
and refiltered with a digital cosine filter. Numerical operating quantities are then calculated from the
processed waveforms.

Filtered data are useful in calculations of fault data. They are the data that the relay uses in the
protection algorithms. Filtered data are not the sampled waveform, but a mathematically manipulated
waveform.

APP751_EventAnalysis_r8 4
 Event Analysis

Waveforms

Copyright © SEL 2019

The raw voltage and current waveforms generated during a power system fault are illustrated on the
slide. These waveforms are the result of a fault on the power system at the end of the line with no
fault resistance. An inception angle of zero gives a full dc offset. The post-fault data are corrupted
with noise, dc, and 2nd, 3rd, and 5th harmonics.

The filtered quantities should look clean without any harmonics, dc offset, or noise. The relay
operates on filtered quantities. Analysis of these quantities will yield proper current and voltage
phasors.

You will want to choose filtered event data for most of your event reporting needs because these are
the signals the relay’s protection elements work from. The raw event reports, however, can also be
useful since this unfiltered data shows what the relay actually “sees” from the power system. Raw
event reports will show conditions such as:

• CT saturation

• Decaying dc offset (the difference between the symmetrical wave and the actual current
wave during a transient condition)

• Power system harmonics (signals with atypical frequencies, creating waveform distortions)

• Input contact bounce

Raw event reports display one extra cycle of data at the beginning of the report.

APP751_EventAnalysis_r8 5
 Event Analysis

Event Report Settings

• ER trigger logic
– Elements other than
trip conditions

– Use rising-edge operators

to prevent long-term
Event Trigger equation
assertion

• LER specifies the length of reports

– Default 15-cycle length allows the relay to store 23 reports

– The 64-cycle length allows the relay to store 7 reports

– WARNING! Changing the LER setting will erase stored event

report data. Save existing data before changing the LER

• PRE specifies prefault cycle length

Copyright © SEL 2019

You can specify the length of event reports as either 15 or 64 cycles of event data in the LER setting.

However, changing the LER setting will erase stored event report data. You must save all existing
event records before changing this setting.

The event report length also affects storage capacity: the default length of 15 cycles allows the relay
to store up to 23 reports. The 64-cycle length allows only 7 reports to be stored due to the higher
resolution and more detailed data capture.

The PRE prefault length setting determines how many cycles of data will precede the event report
trigger point. Prefault data can be helpful in determining what was happening prior to the event.

Elements in the programmable ER logic will trigger event reports for conditions other than trip
conditions, such as external RTD warnings or alarms. An event report is automatically generated
when trip equation elements assert. Thus, any Relay Word bit that causes a trip does not have to be
entered in the ER setting.

It is important to always use the R_TRIG rising-edge operator with all word bits used in the
ER equation. A word bit that remains asserted in the ER equation will not allow future assertions of
the ER equation. It is the assertion of the ER equation that triggers event recording. If the
ER equation remains asserted, no other word bit activity in the ER equation results in event data
capturing.

APP751_EventAnalysis_r8 6
 Event Analysis

Event Report Viewers

1. SYNCHROWAVE® Event 2015

SEL-5601-2015 Software
2. ACSELERATOR Analytic Assistant®
SEL-5601 Software

Copyright © SEL 2019

To view the saved events stored in a compressed format, you will need SYNCHROWAVE® Event
2015 SEL-5601-2015 Software or ACSELERATOR Analytic Assistant Software.

Open an event by double-clicking on a saved event file (.CEV extension)

APP751_EventAnalysis_r8 7
 Event Analysis

SYNCHROWAVE Event 2015 Analysis

Default View

Phase
Currents

Summary
Data
Phase
Voltages

Active Relay
Word Bits

Copyright © SEL 2019

When you open an event report for the first time, the default view includes the three phase currents,
three phase voltages, and active Relay Word bits. Manipulation is required for additional analysis.
 Event Analysis

SYNCHROWAVE Event 2015 Analysis

Split Windows to Add More Charts

Split the chart

vertically

Copyright © SEL 2019

 Event Analysis

SYNCHROWAVE Event 2015 Analysis

Select Different Types of Charts

Copyright © SEL 2019

 Event Analysis

SYNCHROWAVE Event 2015 Analysis

Define Signals in a Chart

Click the gear to

define signals

Copyright © SEL 2019

 Event Analysis

SYNCHROWAVE Event 2015 Analysis

Define Signals in a Chart

Select or search
for signals

Copyright © SEL 2019

 Event Analysis

SYNCHROWAVE Event 2015 Analysis

IA Slide Cursor to Display

Signal Values
IB

IC

VAB
Pin for
VBC Continuous
Display
VCA

Copyright © SEL 2019

Cursors can be slid along the waveform to provide a value for a specific sample point. Multiple
cursors can be added to provide time-difference references. The cursor location values can be pinned
to keep them displayed. Hover over a cursor to display the time between two cursors in seconds or
cycles.
 Event Analysis

SYNCHROWAVE Event 2015 Analysis

Add Tabs for More Space

Add a second page

Copyright © SEL 2019

 Event Analysis

SYNCHROWAVE Event 2015 Analysis

IA_A.Mag

IB_A.Mag Use the Custom Calculation

Engine for to perform
mathematical functions on signals
IC_A.Mag
and graph the results

I2_A.Mag

VBC_V.Mag

VCA_V.Mag

Copyright © SEL 2019

 Event Analysis

Session File Automatically Stores

Analysis Work

Copyright © SEL 2019

All work is stored automatically in a session file in the event report folder. If you close the event and
then re-open it later, the previous analysis work will be preserved. Session files can be centrally
stored and used later as templates to simplify the event analysis for similar relay applications.
 Event Analysis

Search Tools Speed Up Analysis

Analog Channels Settings

Relay Word Bits

Search Contexts
Copyright © SEL 2019

Finding specific event information can be tedious. Powerful search tools help sift through the relay
settings, analog channels, and digital channels to find specific event information quickly.
 Event Analysis

Analysis Process

• Do not guess Understand Expected Operation

• Get the data you Collect Event Data

need Identify Exceptions

• Collect event data Compare to Actual Operation

using the following
event retrieval Determine Root Cause

steps Develop Solution

Copyright © SEL 2019

Before analyzing the details of any event report, start with a basic understanding of what happened or
what should have happened. This generally involves reviewing the relay settings and logic, obtaining
the relay history report, and gathering any other information that may be helpful (known fault
location, targets from other relays, breaker operations, SCADA, and personnel records).

It is important to review normal system operations, in order to more easily recognize problems and
trends of failure.

Rather than guessing what occurred, retrieve the event data provided in the relay, and use the
analysis process to determine root cause.

APP751_EventAnalysis_r8 18
 Event Analysis

Analysis Hint

To determine expected operation, focus on

TRIP and CLOSE outputs and logic
– Find TR logic, decipher elements used, and
research elements

– Do not waste time researching unused

elements

Copyright © SEL 2019

It can be very helpful to obtain a dc control schematic using settings in the protective relay to
understand the use of inputs and outputs. Normally, outputs connected to a breaker will be controlled
by the TRIP and CLOSE logic. These logic equations are functions of elements inside the relay.
Relays contain many of these elements, many of which may not be related to the event. Save time by
researching only the elements that are present in the TRIP logic.

APP751_EventAnalysis_r8 19
 Event Analysis

Example Event
Known Information

• Two SEL-751A Relays: Main and feeder

are fed from an upstream transformer
• Line-to-line fault is suspected (C to A)
• Main relay tripped; feeder relay is not
tripped

Copyright © SEL 2019

In this example event, the customer believed that a line-to-line fault (C-phase-to-A-phase fault)
caused events to be generated in both the main and feeder SEL-751A Relays. A schematic of the
system is provided on the next slide. Upon inspection of the transformer, buswork, and cables, no
fault evidence was found. SEL was contacted to identify what caused the event and why the main
relay tripped the breaker but the feeder relay did not. The event reports from both relays were
provided, along with the settings from each during the time of the event.

APP751_EventAnalysis_r8 20
 Event Analysis

One-Line Diagram
What Happened?
• Main relay tripped the breaker
• Feeder relay did not trip breaker
• No faults were found on the transformer,
buswork, or cables

Copyright © SEL 2019

APP751_EventAnalysis_r8 21
 Event Analysis

Understand Expected Operation

• OUT101 of each relay trips the

appropriate breaker
• Settings in both relays reveal
– OUT101 = TR

– TR = instantaneous overcurrent (50), time-

overcurrent (51), and arc-flash elements
(50PAF, 50NAF, and TOL1 through TOL4)

Copyright © SEL 2019

It is known (or can be observed from system schematics) that OUT101 of each relay trips its
respective breaker. Inspection of the settings inside each relay shows that OUT101 operates upon the
assertion of the TR SELOGIC® control equation, which is a function of instantaneous overcurrent
(50), time-overcurrent (51), and arc-flash elements. The typical installation of arc-flash detection
requires both overcurrent and arc-flash light sensors.

APP751_EventAnalysis_r8 22
 Event Analysis

Understand Expected Operation

Logic diagrams
are available in
the instruction
manual

Copyright © SEL 2019

To understand the operation of the elements in the TR equation, reference the instruction manual.
The instruction manual includes logic diagrams, descriptions, and an explanation of related settings.

APP751_EventAnalysis_r8 23
 Event Analysis

Collect Events and Other Information

Feeder Relay Filtered Event

Is this a fault?
Copyright © SEL 2019

This slide shows the filtered event report from the feeder relay.

Does this look like a fault? It may be hard to tell from this information. Notice the decaying
magnitudes of the phase currents. Also notice that the fault currents are fairly low for a fault
condition. Knowing that this relay is close to a transformer leads to a suspicion of transformer inrush.
You may need to obtain an unfiltered event report to get more information.

APP751_EventAnalysis_r8 24
 Event Analysis

Collect Events and Other Information

Main Relay Unfiltered (Raw) Event

Is this a fault?
Copyright © SEL 2019

This slide shows the raw (unfiltered) event report seen by the main relay. This slide makes it more
obvious that you are looking at transformer inrush instead of a fault.

APP751_EventAnalysis_r8 25
 Event Analysis

Not a Fault

• Decaying dc offset in phase currents

– Harder to see in the filtered (feeder) event
– Easier to see in the unfiltered (main) event

• Transformer inrush
– Neither relay should trip
– Arc-flash overcurrent element is likely to
assert but should not trip without light

Copyright © SEL 2019

Generally, you do not want the relay to trip during transformer inrush. Even more so, you would not
expect an arc-flash element to cause the trip because these elements are typically programmed to
require both overcurrent and light from an arc flash.

APP751_EventAnalysis_r8 26
 Event Analysis

Investigate Settings and Schematics

• Use what you know to work backward
– Breaker opens when OUT101 asserts

– OUT101 asserts when TR asserts

– TR asserts from either 50, 51, or arc-flash elements

 Feeder: TR = ORED50T OR ORED51T OR (50PAF AND TOL1)
OR (50NAF AND TOL1) OR (50PAF AND TOL2) OR (50NAF
AND TOL2)

 Main: TR = ORED50T OR ORED51T OR (50PAF AND TOL1)

OR (50NAF AND TOL1) OR (50PAF AND TOL2) OR (50NAF
AND TOL2) OR (50PAF OR 50NAF AND TOL3) OR IN302

• Use software to view the state of each element

during an event
Copyright © SEL 2019

The breaker opens when OUT101 operates. OUT101 operates when the TR equation asserts, which
is a function of the 50, 51, and arc-flash elements. Look at the logical value of each element in the
TR equation during the event using event analysis software (ACSELERATOR Analytic Assistant) to
determine which element caused the trip.

APP751_EventAnalysis_r8 27
 Event Analysis

Feeder Relay

Copyright © SEL 2019

Starting with the feeder relay, notice that the breaker did not trip because OUT101 did not operate.
OUT101 did not operate because the TR equation did not assert. The TR equation did not assert
because the combination of logical expressions did not evaluate to a logical 1. Although 50PAF did
assert, TOL1 or TOL2 would also have to assert in order to make TR equal to a logical 1.

APP751_EventAnalysis_r8 28
 Event Analysis

Compare Operation to Expectations

Feeder Relay

• 50PAF asserted but not TOL1 or TOL2

– Overcurrent but no light = no trip
– 50PAF = 1, TOL1 = 0, TOL2 = 0
• TR did not assert (expected); would need
50PAF = 1 and either TOL1 = 1 or
TOL2 = 1 to trip for arc flash
• OUT101 did not assert (expected)
• Relay operated as expected
Copyright © SEL 2019

When inspecting the operation of the feeder relay, you can see that the arc-flash overcurrent element
asserted but the light elements did not (as expected). Because the overcurrent elements and light
elements are combined together with AND, the relay did not trip. The feeder relay operated as
expected.

APP751_EventAnalysis_r8 29
 Event Analysis

Main Relay

Copyright © SEL 2019

Observing the event from the main relay, notice that OUT101 operated at the same time the
TR equation asserted. It is also shown that TR asserted at the same time 50PAF asserted.

What is 50PAF? It is an instantaneous overcurrent element for arc flash. Reference the instruction
manual for more information. The instruction manual also explains settings associated with this
element and how to implement it for tripping. Notice that TOL1, TOL2, and TOL3 (the light
elements) are all deasserted during the course of the event. Again, reference the instruction manual
for more information.

APP751_EventAnalysis_r8 30
 Event Analysis

Compare Operation to Expectations

Main Relay
• 50PAF asserted but not TOL1, TOL2, or
TOL3
– Overcurrent but no light = no trip
– 50PAF = 1, TOL1 = 0, TOL2 = 0, TOL3 = 0
• TR asserted; is not supposed to assert
unless overcurrent and light are present
• OUT101 asserted to trip the breaker; this was
expected because TR asserted
• Why did TR assert?
Copyright © SEL 2019

Typical installations require two things to trip for arc flash: a lot of current and a lot of light.
Therefore, if 50PAF asserted, TOL1, TOL2, or TOL3 would also have to be asserted. This was not
true for the event. To find out why, inspect the TR equation in the main relay.

APP751_EventAnalysis_r8 31
 Event Analysis

Use the Instruction Manual to Find

Root Cause
• Verify logic
equations
• Note order of
operations:
AND comes
before OR

Copyright © SEL 2019

First, review the information on SELOGIC control equations in the instruction manual, such as the
table shown on this slide that indicates the order of operations.

APP751_EventAnalysis_r8 32
 Event Analysis

Root Cause Found

• Main relay: Mistake was made

– TR = …OR (50PAF OR 50NAF AND TOL3) OR…
 Relay reads this as 50PAF OR (50NAF AND TOL3)

 Intention was to have (50PAF OR 50NAF)

AND TOL3

– TR asserts when 50PAF asserts; the presence of

light is not necessary

• Feeder relay: Mistake was not made

Copyright © SEL 2019

This slide shows one section of the TR equation in the main relay. The user who entered the
expression had a different interpretation of the expression than the relay did because ANDs are
evaluated before ORs. Therefore, the presence of light was not necessary, and the relay tripped the
breaker as soon as 50PAF asserted.

APP751_EventAnalysis_r8 33
 Event Analysis

Benefits of Analysis

• Understand the system better

• Verify correct relay operation and the
cause of a misoperation
• Identify importance of proper testing
during commissioning

Copyright © SEL 2019

After finding root cause, the solution should be implemented and tested. Through the process of
event analysis and troubleshooting, you can get a better understanding of your system, verify relay
operation, find the cause of a misoperation, and correct settings or wiring errors when necessary.
Proper testing during commissioning is important and can help reduce the frequency of unexpected
operations, as well as limit the amount of time necessary to determine root cause.

APP751_EventAnalysis_r8 34
 Event Analysis

Five Things to Remember About

Event Data
1. The event report of interest may not be
the newest event
2. The relay does not need to trip to
generate an event report
Use the ER equation to trigger non-trip events

The relay generates an event

report on the rising edge of
the phase or ground
time-overcurrent element.

Copyright © SEL 2019

The ER Trigger equations use the R_TRIG logic operator, which will momentarily assert an element
on its rising edge.

APP751_EventAnalysis_r8 35
 Event Analysis

Five Things to Remember About

Event Data
3. Download event data soon: Event report
storage is limited
4. Download both filtered and raw data
– Filtered is best for standard protection
analysis

– Raw is needed to see inrush, CT saturation,

and harmonics

Copyright © SEL 2019

APP751_EventAnalysis_r8 36
 Event Analysis

Five Things to Remember About

Event Data
5. Event retrieval of combined or correlated
relay events requires highly accurate time
synchronization
– System-wide event analysis is greatly simplified
with time synchronization
– Connect a relay time source:
 IRIG-B input signal

 GPS clock

Copyright © SEL 2019

One-millisecond resolution is required to accurately align event data. Most alternate time-
synchronization methods (manual relay, computer clock settings, or SCADA protocols) do not meet
event data time accuracy requirements.

The SEL-751 has three different physical interfaces to provide demodulated IRIG-B time-code input
for time synchronization. If the relay has multiple options for IRIG-B input, only one input can be
used at a time. Refer to the SEL-751 Instruction Manual for more information.

For more information on event report analysis, see the technical paper “Forward to the Basics:
Selected Topics in Distribution Protection” by Lee Underwood and David Costello, on your Student
CD and available at selinc.com.

APP751_EventAnalysis_r8 37
 Event Analysis

Generator Event

Copyright © SEL 2019

APP751_EventAnalysis_r8 38
 Event Analysis

Mystery Event

Copyright © SEL 2019

APP751_EventAnalysis_r8 39
 Event Analysis

Mystery Event

Copyright © SEL 2019

APP751_EventAnalysis_r8 40