Fuzzy Cognitive Map for Health Assessment of IEC 61850 ... - aceps

Fuzzy Cognitive Map for Health Assessment of IEC
61850 Based Devices and Functions
Salman Mohagheghi
ABB Inc., US Corporate Research Center, Raleigh, NC, USA
© ABB Group
March 3, 2011 | Slide 1

Contents
Introduction
Communication within the Substation
Overview of IEC 61850
Communication Services
Data Model
Health State of Devices and Functions
Cognitive Maps
Historical Approach
Fuzzy Logic
Fuzzy Cognitive Maps
Case Studies
Concluding Remarks
© ABB Group
March 3, 2011 | Slide 2

Introduction
Smart Grid
Communication and metering infrastructure
Self-healing mechanisms
Desired power quality level
Advanced monitoring systems
Efficient demand responsive programs
Future proof design
Interoperable infrastructure
© ABB Group
March 3, 2011 | Slide 3

Communication within the Substations
IEC 61850 Overview
IEC 61850 is a standard specifically designed for implementing
substation automation (SA) systems.
The main features of the standard are:
Data Modeling- Object oriented platform-independent
modeling of all the functions.
Fast Transfer of Events and Messages- GOOSE and
GSSE for the peer-to-peer communication mode.
Setting Groups- Control blocks allow the user to switch
from any active group of setting values to another.
Sampled Values- Transfer of sampled values from modern
switchgear and nonconventional CT/VTs.
Data Storage and Representation- Configured data of the
substation is stored in XML format using the Substation
Configuration Language (SCL).
© ABB Group
March 3, 2011 | Slide 4

Communication within the Substations
IEC 61850 Substation Automation Topology
© ABB Group
March 3, 2011 | Slide 5

Communication within the Substations
Services
© ABB Group
March 3, 2011 | Slide 6

Communication within the Substations
Logical Node
Virtual Units: objects defined in
the object oriented context of
the standard.
LN is a collection of data
objects, data-set objects,
descriptive attributes, report
control objects, log-control
objects and a list of sampled
values.
The type of data contained in a
LN can be operational (e.g.
measurement values, position
status) or configuration data
(e.g. IED self descriptive).
The substation automation
functions are broken down into
LNs.
© ABB Group
March 3, 2011 | Slide 7

Communication within the Substations
IEC 61850 Data Models
© ABB Group
March 3, 2011 | Slide 8

Communication within the Substations
Health State in IEC 61850
IEC 61850 defines health attributes for the logical nodes and
the external devices connected to them. These include:
Health: state of the logical node related to hardware or
software. Mandatory data with three states:
Ok (green): no problems, normal operation
Warning (yellow): minor problems, but in safe operation
mode
Alarm (red): severe problem, no operation possible
EEHealth: state of the external equipment, for example a
circuit breaker controlled by the logical node XCBR.
Optional data with the above three states.
PhyHealth: Mandatory attribute that refers to the health
state of the physical device.
© ABB Group
March 3, 2011 | Slide 9

Cognitive Maps
Historical Approach
Introduced in 1970s for representing social scientific
knowledge.
Signed digraphs where a positive link from vertex A to vertex
B indicates that A causally increases B, whereas a negative
link from A to B means that should A happen, the chances of
B happening are reduced.
Indirect Effect of A on B over the path pi is negative if the
number of negative signs (of the edges) is odd, and
positive otherwise.
Total Effect of A on B would then be derived from the
indirect effects over all possible paths: the total effect of
A on B is negative if all the indirect effects are negative,
is positive if all the indirect effects are positive, and
indeterminate otherwise.
© ABB Group
March 3, 2011 | Slide 10

Cognitive Maps
Weighted Approach
Alternative Solution: creating a weighted digraph
removes the uncertainty related to the sign calculations
requires higher level of granularity in defining the
dependencies of causes on one another
w
AB p
= wA
, j
× w
j , j
× L w
j , B
,
×
i
1 1 2
w
AB
wAB,
p
= ∑
i
i
n
© ABB Group
March 3, 2011 | Slide 11

Cognitive Maps
Fuzzy Logic
Causality is fuzzy
Fuzzy logic can deal with imprecise and vague information. It
enables representation of human knowledge with
approximate terms and values
© ABB Group
March 3, 2011 | Slide 12

Fuzzy Cognitive Maps
Qualitative Approach
FCM provides a qualitative output (not a crisp number)
w
w
AB
AB
= min { wA,
j
, w
j , j
, L,
j ,
, p
w
i 1 1 2
n B
m
= max { wAB
p
, L,
wAB,
, 1 pm
}
}
In other words:
Indirect Effect: weakest causal link in a path
Total Effect: Strongest of the weakest links
© ABB Group
March 3, 2011 | Slide 13

Case Study
Transformer Bay
Transformer bay with protection/control functions:
line distance protection, tap changing functions with related protection
functions, interlocking and reclosing functions, general monitoring
functions, and a generic control function for regulating the transformer
tap positions
© ABB Group
March 3, 2011 | Slide 14

Case Study
Fuzzy Cognitive Map
© ABB Group
March 3, 2011 | Slide 15

Case Study
Voltage Transformer
© ABB Group
March 3, 2011 | Slide 16
TVTR1: “Alarm” state for health.
The impact of this health status on the IHMI (at the operator
desk):
TVTR1 → MMXU → IHMI
TVTR1 → PTOV → AVCO → GAPC → IHMI
TVTR1 → CPOW → GAPC → IHMI
TVTR1 → MMXN → GAPC → IHMI
TVTR1 → PDIS → RREC → IHMI
The total effect of the Alarm health status of the TVTR1 on the
IHMI is determined as:
max{ min(V.H, V.H) + min(H, M, H, M) + min(L, H, M)
+ min(H, M, M) + min(V.H, V.H, V.H) }
= Very High

Case Study
Circuit Breaker
CB controlled by XCBR: “Warning” state for health (i.e. Warning
state for the EEHealth data of XCBR)
The impact of this health status on the performance of the
generic GAPC control function:
XCBR → GAPC
XCBR → CILO (Bay) → CILO (Station) → GAPC
XCBR → CSWI → CILO (Bay) → CILO (St.) → GAPC
The total effect of the Warning health status of the CB on the
GAPC is determined as:
max{ min(M) + min(V.H, M, M) + min(V.H, V.H, M, M) }
= Medium
© ABB Group
March 3, 2011 | Slide 17

Case Study
Tap Changer
YLTC: “Alarm” state for health
The impact of this health status on the performance of the
ATCC logical node:
YLTC → ATCC
YLTC → AVCO → GAPC → ATCC
The total effect of the Alarm health status of YLTC on the
ATCC is determined as:
max{ min(V.H) + min(H, H, V.H) } = Very High
© ABB Group
March 3, 2011 | Slide 18

Concluding Remarks
IEC 61850 standard defines a data attribute indicating the health
status of logical nodes (i.e., basic functions) and physical devices.
This attribute tags the data points traveling across the system.
While this identifier attribute can be easily interpreted and utilized
in the device immediately connected to the source of the data, it is
often not easy to find the impact of a non-ideal health status in a
web of multiple control and protection functions.
A fuzzy cognitive map (FCM) is proposed for evaluating the impact
of non-ideal health status of logical nodes and physical devices in
an interconnected network on the performance of the rest of the
system functions.
FCM can monitor the devices and functions with a predefined
frequency to raise flags upon detecting non-ideal health.
Expert knowledge can be used to determine the weights of the
interdependences of logical nodes on one another.
© ABB Group
March 3, 2011 | Slide 19

© ABB Group
March 3, 2011 | Slide 20