Condition-Based Maintenance

You might also like

You are on page 1of 4

Condition-based Maintenance: A Forward Step on the Power

System Protection Maintenance at the Portuguese Transmission


System Operator
H Leite*, B Soares*, S Costa †

* University of Porto-Faculty of Engineering, Portugal, hleite@fe.up.pt and ee08236@fe.up.pt


†ELEX-ES – REN, Portugal, susana.costa@ren.pt

Based on the protection system condition (named Key


Keywords: Protective Relays, Condition Based Maintenance, Performance Indicators (KPI)), it is assessed the protection
Key Performance Indicators, Disturbance Records. system assets to predict its risk of failure in real-time.

Abstract 1.1 The ¶V Protection system Maintenance Policy at the


Portuguese TSO
This paper demonstrates an advanced model which integrates
both corrective and condition-based maintenance in the Every power equipment at the network is being monitored and
Portuguese Transmission System Operator (TSO) protection protected by two different and competing protection systems.
system. It also identifies the relevance of protection system This protection systems usually comprises current and voltage
condition indices for the maintenance decision-making transformers, protective relays, batteries, circuit breakers and
process. The developed model analyses data from every fault communications channels.
event till a three level back-up system and decides which type
of maintenance to apply. A detailed example of this decision- Figure 1 illustrates the protection system redundancy.
making model is illustrated in this paper.

1 Introduction

Power protection system has the duty to protect personnel and


equipment as well as to limit damage in the case of electrical
faults. Whenever necessary, protection system needs to operate
correctly and in a highly reliable manner regardless of being
idle for months or years. To achieve a high level of reliability,
effort in maintaining equipment over its lifecycle is needed.
Different maintenance policies can be applied. There have
been various classifications of maintenance policies and they
can be categorised into reactive (or corrective) maintenance, Figure 1 Schematic of the protection V\VWHP¶ redundancy
adaptive maintenance, preventive maintenance and condition-
With this redundancy in mind, the Operation and Maintenance
based maintenance (CBM).
Division is comfortable applying corrective or adaptive
maintenance to the protection systems instead of any other
Due to the fact that Portuguese TSO has a redundancy over
maintenance scheme e.g. Time Based Maintenance. However,
each protection system (i.e. each primary equipment is
at the moment, all the protection systems components are
protected by two different and competing protection systems
monitored and analysed after a network failure. It is the way to
with different voltage and current transducers, protective
know if any corrective maintenance action is required. This
relays and protective algorithms, different relays source supply
reactive maintenance can be realized either locally or remotely.
and also different circuit breakers trip coils), the maintenance
policy so far is mainly reactive. However, aiming to seek
Independently of the protection system worked correctly, there
higher reliability, the Portuguese TSO is committed to apply
is always a health check of the primary and backup protection
condition-based maintenance policy to its protection system.
system. In the Portuguese Transmission Network between 300
Thus, the Portuguese TSO intends to prioritise and optimise
and 400 incidents take place in average each year; most of them
maintenance resources based on the real-time data. This work
due to external causes such as forest fires, storms, pollutions,
advances a model which integrates both corrective and
birds or lighting.
condition-based maintenance to the protection system.

Authorized licensed use limited to: DNV GL Library. Downloaded on June 03,2020 at 07:01:19 UTC from IEEE Xplore. Restrictions apply.
2 Network Failure: Seeking for a Protection 3 ³&HQWUDO Application of Disturbance
System Automatic Analysis $QDO\VLV´ : The Decision-making Model
Protection systems fail if: (i) primary protection (i.e.,
protection whose tripping is necessary to eliminate the fault) The application analyses data after a network failure event in a
does not work properly; (ii) fault eliminated by backup three level of back-up system and decides which type of
protection; (iii) the fault elimination time is greater than a maintenance to apply. The application is therefore a decision-
maximum value allowed; (iv) protection tripped in untimely making model as shown in Figure 2. The type of maintenance
way; (v) fast autorecloser GRHVQ¶W work properly; (vi) circuit devised is either corrective, adaptive or condition-based
breaker GRHVQ¶W work properly. From all these functional maintenance and is based on the component KPI¶V as
failures, it is possible to determinate failure modes. previously mentioned.
Considering numerical protections, failure modes are: (i)
failure in alimentation system; (ii) problems in protection
HTXLSPHQW¶V algorithm; (iii) mechanical contacts of
protectLRQ¶V exit signals jam; (iv) analogic-digital converter
and input cards with problems; (v) problems in input/output
cards of digital signals; (vi) problems in memory or other
electronic parts of protection equipment; (vii) problems in
GHYLFHV¶ timers; (viii) incorrect parameters/settings; (ix)
protection DOJRULWKP¶V bugs; (x) problems in windings of &7¶V
and 97¶V (xi) problems in current EUHDNHUV¶ tripping coils. All
these failure modes must be observed when defining a
maintenance strategy to power systems relaying.

The rationale behind the development of the ³&HQWUDO


Application of Disturbance $QDO\VLV´ is to acquire and analyse
every network failure automatically and to decide which type
of maintenance should be applied. Thus, this application takes
over all the systematic tasks, update all protection system
condition indices and generates maintenance alarms in real-
time. These protection system components condition indices
(named Key Performed Indicators, KPIs) could be intrinsic to
that particular protection system component or common to that
protection system component family.

Figure 2 ³Central Application of disturbance Analysis´


2.1 Key Performance Indicators .3,¶V  Specific & flowchart
Common
The Condition Based Module combines two sub modules, i.e.
the ³&RQGLtion Monitoring 0RGXOH´ and ³Decision-making
Two sets of Key Performance Indicators are derived to help Module´. The condition monitoring module task is to
indicating the condition of every protection system component calculate, to update the KPI¶s and store all indicators values in
health. These two sets are divided in those who belong only to the database. The decision-making module task is to determine
that particular component and those that may be common to the equipment¶s alarm level in terms of maintenance. The
that component family (i.e. make and model) alarm level can be: (0) ± no alarm; (I) ± Maintenance action
should be taken; (II) ± Maintenance action is required;
The Key Performance Indicators are essential to automatically Additionally to the alarm level issued to the operator, a report
and in real-time analysed the protection primary and back with the critical KPI¶s values is created and sent to the operator
system´s components condition and therefore help the ³FHQWUDO as well. If a maintenance action is required and taken, the KPI¶s
Application of disturbance Analysis ³ to decide where there will be updated accordingly.
will be a maintenance action or not, if so, which type of
maintenance action should be more appropriate to that 4 A Working Example
component. The .3,¶V are also divided into classes depending
on which protection system component is being assessed. In order to demonstrate how the importance of network
disturbance is identified, and Key Performance Indicators
.3,¶V) specifics and commons are determined in the
maintenance action decision-making, a working example is
used in the following.

Authorized licensed use limited to: DNV GL Library. Downloaded on June 03,2020 at 07:01:19 UTC from IEEE Xplore. Restrictions apply.
An incident happens between the 400kV substation SFF and From the chronological event and oscillography it is possible
PCRJ and the overhead line was tripped off. to make the following decisions: i) whether or not the fault is
in the trip zone ii) which phase(s) are faulted. iii) which
protective function(s) are responsible to open the circuit
breaker iv) whether or not teleprotection works v) if the backup
in adjacent substation sees the incident and have the same
perception about the type of fault and which protective
functions are activated vi) whether there is a single pole
tripping and/or reclosing, and so on.
From the oscillography synchronized time it is necessary to
find its location in order to make necessary repairs and restore
power as soon as possible. An accurate fault location on the
line is an important requirement both for a permanent fault and
for a transient fault that may point to a weak spot.

Figure 3 the 400kV network with the indication of the incident


between substations PCRJ and SFF

The analysis of the incident starts with reading the


chronological event from both substations. The chronological Figure 6 graph locating the fault
event must be with the synchronized time. The oscillography
will be read and carefully analysed.
As the fault distance from both substations was found and as
the distance relay was active, it is important to plot the
trajectory of the pre-fault, fault and post-fault on the R-X plane
from both distance relays.
R-X plane explains sets of relays for protection of each line.
Each relays provides characteristic for each zone. It is possible
to find if the relay will see the fault as beyond its characteristic
and will not operate even though it should have operated, this
is called ³XQGHUUHDFK´ or instead ³RYHUUHDFK´ i.e., to operate
for a larger value of impedance than that for which it is adjusted
to operate under steady-state conditions.

Figure 4 the oscillography obtained from both substations

Figure 6 The R-X diagram from both ends of the overhead line.

Now all this goes to the ³&RQGLWLRQ Monitoring 0RGXOH´


which in turn is going to update all the ³.H\ Performance
,QGLFDWRUV´ specific and/or common.
Figure 5 Time of the events occurred with the active protective
functions

Authorized licensed use limited to: DNV GL Library. Downloaded on June 03,2020 at 07:01:19 UTC from IEEE Xplore. Restrictions apply.
The Key Performance Indicators Specific to the failure are References
advanced and determined. Particular to the IED itself
(temperature, humidity, operating times, number of years after [1] (in Portuguese) A. Marques, Análise Comportamental de
last maintenance), operating time, non selective trip, operating Sistemas - Sistemas de Proteção - Critérios de classificação dos
time excessive, correct operating time, incorrect operating defeitos para preenchimento de base de dados da aplicação
time, number of time failure to pickup, failure to operator due Gestcomp, Documentos Internos ed.: REN - Redes Energéticas
to saturation of the current transformer; failure due to the Nacionais 2008.
secondary circuit of the protection system, .3,¶V could also be [2] J.-H. Shin and H.-B. Jun, "On condition based maintenance
related to the incident itself as: protective relay operation time, policy," Journal of Computational Design and Engineering,
time of the protective relay pick-up; number of relay pick-up vol. 2, pp. 119-127, 2015.
well-succeed, number of single pole reclosing, circuit breaker [3] Tavares, H., et al. Applying Reliability Centered
opening time, etc. Maintenance to a digital protective relay. in Innovative Smart
Grid Technologies (ISGT Europe), 2012 3rd IEEE PES
The .3,¶V common to a family of protection system International Conference and Exhibition on. 2012.
components might be: watchdog failures, mean time between [4] A. A. M. Zin and S. P. Abdul Karim, "The utilization of
failures, non selective failures by IED model and software digital fault recorders in protection system analysis on Tenaga
version, number of operating time excessive, number of time Nasional Berhad transmission system," Power Delivery, IEEE
failure to pickup, etc. Transactions on, vol. 22, pp. 2040-2046, 2007.
[5] (in Portuguese) REN, Relatório de Qualidade de Serviço
Final Remarks 2014: REN, 2015.
[6] J. Moubray, "Reliability Centered Maintenance," ed
With the work presented, one can conclude that KPIs allows Amsterdam: Elsevier Butterworth-Heinemann, 1997.
engineers to know when and where there are problems in [7] J. H. Spare, "Building the business case for condition-based
protection power systems. It was presented the Condition maintenance," in Transmission and Distribution Conference
Based and Corrective maintenance application. Also, with this and Exposition, 2001 IEEE/PES, 2001, pp. 954-956 vol.2.
implementation, an operating equipment historic (individual [8] D. Zhao, W. Liang, and X. Zhang, "Relay protection
and by model) is built. This allows future implementation of condition assessment based on variable weight fuzzy synthetic
reliability and risk based maintenance. Without relevant evaluation," in Technologies for Sustainability (SusTech),
reliability data, this model seems to be the most adequate to be 2014 IEEE Conference on, 2014, pp. 115-120.
implemented in Portuguese TSO. [9] L. Zhijun, D. Zhihui, and J. Yanjun, "Research on
For future development, economic viability about PRGHO¶V condition-based maintenance strategy of protection systems
implementation in Portuguese TSO can be examined and the based on multistage fuzzy comprehensive evaluation," in
goal of Decision Taking Module development can be further Fuzzy Systems and Knowledge Discovery (FSKD), 2014 11th
investigated. International Conference on, 2014, pp. 268-272.
[10] P.-C. Lin, J.-C. Gu, and M.-T. Yang, "An intelligent
maintenance model to assess the condition-based maintenance
of circuit breakers," International Transactions on Electrical
Acknowledgements Energy Systems, 2014.
.
The authors are grateful to the Operation and Maintenance
Division from the Portuguese Transmission System Operator,
Redes Energéticas Nacionais, REN for the collaborative work.

Authorized licensed use limited to: DNV GL Library. Downloaded on June 03,2020 at 07:01:19 UTC from IEEE Xplore. Restrictions apply.

You might also like