Protective Control and Fault Diagnosis of High-speed Railway ...

The 35th Chinese Control Conference (CCC)

Protective Control and Fault Diagnosis of High-speed Railway

Traction Power System

Reported by: QIAN Qing-Quan

Southwest Jiaotong University

2Layout

11• Background and Traction Power System

22• Automation for High-speed Rail Traction

Substation

33• Detection and Diagnosis of Pantograph -

Catenary System

44• Prognostics and Health Management of

High-speed Rail Traction Substation

3

Background and Traction Power SystemPart I

41. Background and Traction power systemAt the end of 2015, the operating mileage of CRH is about19,000 km; which will reach to 30,000km at 2020.

1203300

19000

30000

2008 2009 2015 2020CRH Operating Mileage (km)

Two major problems of CRH:Two major problems of CRH:1. Scientific construction1. Scientific construction

2. Safety and high-efficient operation2. Safety and high-efficient operation

The 35th Chinese Control Conference

5CRH 3.0 “8 vertical 8 horizontal" railway net

1.0 “4 vertical 4 horizontal”2004~2008

2.0 “Long-term railway net plan(2008)”

3.0 “8 vertical 8 horizontal”2016~2030

6

Traction power supply (TPS) system is the only way toprovide power for the train as well as other systems.

1. Background and Traction power system

Maintenance syst.High-speed train Commun. syst.

220kV Grid Traction substation Catenary

Service syst.


7

Jing-Hu high-speed rail (Shanghai Bureau)

Sections Faults in 2013 Faults in 2014 Fault Ratio

Suzhou Maintain 37 times 24 times 3.9 times / hundred km*yearBengbu Maintain 19 times 22 times

Adverse affect

Short circuit

Traffic delay

Acc

iden

ts

Operation Environment; Load Behavior; Special Structure

Fault ratio

1. Background and Traction power system Fast & accurate fault remove, power restore and reliable

operations of TPS are urgent needed

81. Background and Traction power systemTPS consists of Traction substation, Catenary system andLocomotive tractive system.

Automation of traction substation

Monitoring of Catenary-pantograph system

Physical health diagnosis management (PHM)

Protection and control; fault location;electrical conditions monitoring;

Monitoring catenary-pantograph system operations, quickly find the faults of components

Power grid CatenaryTraction substation High-speed Train

220kV 27.5kV 25kVPantograph

Fast fault diagnosis; Health assessment; Reliability analysis; Risk assessment; Maintenance decision

Our

wor

k

9

Automation for Traction SubstationPart II

10

High speed：> 300km/h

High density：3-5 min

Long group：16 unit 20.4MW

Direct /BT-fed catenary

Ynd11 transformer

110kV power source

Rail/ return flow

(China high-speed railway) CRH

Foreign high speed railway

Direct /AT-fed Catenary

Single phase /Scott traction transformer

121~154kV power source

Integrated grounding system

Conventional speed railway

All parallel AT-fed catenary

V/X traction transformer

220kV power source

Integrated grounding system

2. Automation for traction substation


It is a new challenge for the novel traction modes

11

Impedance‐matching balance Y/△变 V/V） Single‐

phase SCOTT V/X Y/△/△

V/V transformer Single-phase transformer

Yd11 Transformer

Scott transformer

Cross transformerImpedance matchingbalance transformer

V/X Transformer

Due to different connections of traction transformer, a unified protection model has been presented in

terms of different matrices shown in the following table.

Transformer protection technology

1 2

1 2

1 2

A A

ABC B B

C C

K KI K K I

K K

ACT

CT

nKn

_

_

**1

_

_*CT

CT B

nK n

CCT

CT

nKn

_

_

**1

CCT

CT

nKn

_

_

**1

_

_*CT

CT A

nK n

BCT

CT

nKn

_

_

**1

BCT

CT

nKn

_

_

**1

CCT

CT

nKn

_

_

*

ACT

CT

nKn

_

_

*2

ACT

CT

nKn

_

_

*2*732.1

BCT

CT

nKn

_

_

*2

BCT

CT

nKn

_

_

*2*732.1

CCT

CT

nKn

_

_

**1

0

0

0 0

0

0

0

ACT

CT

nKn

_

_

*

BCT

CT

nKn

_

_

**1

ACT

CT

nKn

_

_

*

_

_

1**

CT

CT B

nK n

BCT

CT

nKn

_

_

**1

CCT

CT

nKn

_

_

*

ACT

CT

nKn

_

_

*

BCTnKn

_*

CCT

CT

nKn

_

_

**1

CCT

CT

nKn

_

_

**1

ACT

CT

nKn

_

_

**366.1

ACT

CT

nKn

_

_

**366.0

BCT

CT

nKn

_

_

**366.0

BCT

CT

nKn

_

_

**366.1

CCT

CT

nKn

_

_

**1

CCT

CT

nKn

_

_

**1

1AK

2AK

1BK

2BK

1CK

2CK

0

0

0

0

0

0



12

upstream

downstream

Z<I>

△I>AR

U<AR

AT Section post

Z<I>

△I>AR

U<AR

U<AR

U<AR

Traction substation

Fast tripping Pressure loss solution

Recovery of power supply

With electricity

Without electricity

Power supply arm

Protection and control: the cooperative protection principle of the all parallel AT-fed traction arm is proposed.


Up track

Down track


13

0 1 2max( , , )g g g setI I I I

T1

AT1

AT Section post

AT2

T2

F2

F1

R2

R1

01IT

01IF

02IF

11FI

02IT

11TI

12TI12FI

21FI21TI

22TI 22FI0gI

1gI2gI

Upstream

Downstream

Transformer

substation

TR fault

TF fault

0 1 2max( , , )g g g setI I I I

1 1 2 2T F T FI I I I

1 130 arg( , ) 150T FU I I

1 1 2 2T F T FI I I I

1 1150 arg( , ) 330T FU I I

Horizontal line current maximum DiscriminantTraction substation AT、Section post stream type

down TF

up TF

Horizontal line current maximum Discriminant

IT1 max IF1max IT2max IF2max Up and down

Type

√ down TR

√ down FR

√ up TR

√ up FR

It breaks through the fault identification technology of AT traction network, and realizes the

fast and accurate identification of fault direction and fault types.

Traction network fault identification

Fully parallel AT catenary


14

0 2000 4000 6000 8000 10000f/Hz

0

50

100

幅值

Primary natural frequency fs

The fault wave reflects between fault point

and measuring device.

Additionally, because of the delayed effects

of the transmission line with finite length,

the fault traveling wave forms a spectrum

which consists of a series of harmonic

frequencies. It is called the natural

frequency of the fault traveling wave.

( 2 ) 0, 1, 2...4

S FS

k vf kd

，

We derived the relationship among natural frequency, and fault location and boundary condition is,

14/79页



15

信息

Old way > 20 ms

Fundamental

New way

Transient

< 10 ms


The natural frequency method is used to present a fast frequency-domain protection theory together with transient information from fault.


16

3. Promote the technique improvementEstablished 9 technique standards, in which 5 standards have been authorized.


No Code Title

1 GB/T18038-2008 General specification of microprocessor-based protection equipment for electrified railway traction power supply system

2 TB10117-2008 Code for design of railway power supply dispatching system

3 TB/T3226-2010 Integrated automation system and devices of traction power supply substation for electrified railway

4 运电通信函2012-248 Technical scheme of communication network for railway power supply dispatching system

5 铁运函2012-136 Technical specification for high speed railway power supply safety monitoring system

6 Integrated SCADA system for power supply of Passenger Dedicated Line -Standard for man-machine interface drawing of dispatching system

7 Integrated SCADA system for power supply of Passenger Dedicated Line -Table specification

8 Integrated SCADA system for power supply of Passenger Dedicated Line -IP address planning criteria

9 Integrated SCADA system for power supply of Passenger Dedicated Line -104 statute

17

① Power supply dispatching system

④ Remote terminal unit

RTU

② Traction substation automation

system

③Power distribution automation system

Controlled station system

Equipment development：First developed aintegrated monitoring system for high speed railway power supply in China



18

Power supply dispatching system

Dispatching

station

Ji'nan Rail Bureau

Guangzhou Railway Group

BeijingRail Bureau

Shenyang Rail Bureau

武汉Harbin Rail Bureau

Shanghai Rail Bureau

Wuhan Rail Bureau

Application

software

Cloud comput

ing platfor

mZhengzhou Rail Bureau



19

Automatic traction substation system: 17 devices



20

Electric substation automation system（16devices contained）



21

Developed 5 categories of RTU

GM7-DGM7-CGM7-BGM7-AGM7-OPU



22

Application Promotion in CRH



23

Detection and Diagnosis of Pantograph - Catenary SystemPart III

243. Detection and Diagnosis of Pantograph - Catenary System

19000KM

Status

Safety technologyUrgently Needed!2005 2009 2010 20112008 2012 2013 2014 2015

兰新厦

深京广

京沪

郑西武

广京津

合福

China High-speed railway（before2015）

Pantograph-catenary system


25

￭ Detect faults of Pantograph-catenary effectively?￭ Diagnose faults of Pantograph-catenary early?￭ Determine strategy of maintenance scientifically?

￭ Detect faults of Pantograph-catenary effectively?￭ Diagnose faults of Pantograph-catenary early?￭ Determine strategy of maintenance scientifically?

Conventional methods Developing directions

■ patrol on foot ■ static detection◆ sensor

◆ communication

◆ network

◆ information

◆ computer

■ patrol on platforms ■ dynamic detection

■ patrol in vehicles ■ overall detection

■ patrol by EMU ■ un-regular detection



26

1C 2C 3C

4C 5C 6C

Comprehensive data center

1C: Comprehensive pantograph and catenary monitor device2C: Catenary-checking video monitor device3C: Catenary-checking online monitor device4C: High-precision catenary-checking monitor device5C: Pantograph video checking device6C: Ground monitor device for catenary and power supply equipmentData center

6C

Hierarchical and distributed system



27

TheoryMachine vision

ApplicationGeometric parameter

Camera calibration for high-precisionExtracting light stripe centerImage processing

￭ Height of contact wire￭ Zig-zag value of contact wire

H

L

A B

1C

1a

2a3a

4a

5a

6a

7a

8a

4C

7C

5C

6C

L拉出值

D导高

2C3C

wx

wy

wo

wz

8C

9C

Critical technology 1



28

TheoryMulti-sensor Information

Fusion

ApplicationStrategy based on arcs, force,

displacement, temperature, etc.Ultraviolet sensor Infrared sensorData fusion: Bayes, D-SDeep learning algorithm

￭ Detection￭ Comprehensive evaluation

240 250 260 270 280 290

0

1

2

3

4

5

6

7

8

9

10

x 108

波长（nm）

光谱

响应

度（

mV

/W）

光谱响应度曲线




29

TheoryImage processing

ApplicationComponents malfunction

Image enhancementFeatures extractionTemplate matchingClassifier

￭ quickly find the faults of components




30

TheoryBig data and cloud computing

Applicationmanagement

Prognosis and health managementVirtualization Technology Distributed parallel architecture

Centralized analysis, monitor, management




31

Definition

1CFixed detection device is installed on high speed comprehensivetrain. With the operation of comprehensive train，the detectionresults which include security and geometric parameters are usedto guide the repairs of the catenary.

ObjectiveThe target of detection is that making clear of the statement of thecatenary-pantograph, including the geometric parameters ofcatenary, wear, arcs, etc.

Drafted by SWJTUPromulgated by the Ministry of Railways



32

Definition

2CPortable video device for detecting the state of catenary intemporary installation, to carry on statistics and analysis tothe status of catenary by image and guide the maintenance ofcatenary condition.

ObjectiveDetection target is the significant changes of catenarytechnology status , such as nest, tree damage, contactsuspension changes etc., recorded by video.


Dropper missed



33

Definition

3CUnder condition of usual running, the device installed on the top ofEMU monitors the state of catenary and pantograph, achieve fullcoverage and dynamic detection.

ObjectiveTarget detection for dynamic geometric parameters includingheight, zigzag value, the distance between the two lines,pantograph catenary arc (arc rate, duration and arc energy, arcnumber etc.) and temperature of contact suspension etc.


0 2 4 6 8 10 12 14 16

x 104

0

100

200

300

400

时间（单位ms）

紫外燃弧能量（单位

mJ）

匀速340Km/h

0 2 4 6 8 10 12 14 16

x 104

0

100

200

300

400

紫外燃弧能量（单位

mJ）

匀速330Km/h



34

Definition

4C

Under a certain speed, using operating or special vehicles ascarrier, capture the images of catenary and its components, andmeasure the static catenary’s geometric parameters, andmaintenance suggestions are formulated through the data andimage analysis.

ObjectiveThe object is to obtain geometric parameters (including the height ofcontact wire, zig-zag value, the distance between the two lines), andfind quicly mechanical failure (loose, fracture, missed, crack,damaged components).




35

Definition

5CAt the boundary of bureau, power supply department or at throatarea, to capture images of passing pantograph in order tomonitor the abnormal state of carbon, and ensure the safety ofthe catenary-pantograph.

ObjectiveTo acquire the image of pantograph slide surface, and check theabnormal situations including deformity, crack, excessive wear.


Burn



36

Definition

6CBe set up in the special section of catenary and the tractionsubstation, and to carry out the real-time monitoring of the singleparameter of catenary or the power supply equipment in order toguide the maintenance.

ObjectiveThe monitoring objects include the tension force, vibration, uplift,temperature, compensation displacement of catenary and theinsulation of the power supply equipment, cable joint temperature.


0 50 100 150 200 250 300 350 400-50

0

50复合13号杆绝缘子

采样点/个

泄漏

电流

/mA



37

China Railway CorporationData center

Railway BureauData center

Power supply departmentData center



38

Prognostics and Health Management(PHM) for traction substationPart IV

39

Online monitoring device

Offline test

Sensor networksSCADA system

Test vehicle

Protect Device

Environmental information

Multi source databaseData source of traction Power supply equipment Signal processing

method

Fault prognosis

Fast fault diagnosis

Health assessment

Reliability analysis

Short time scaleMedium time scaleLong time scale

Risk assessment

Residual life prediction

Cost

Time Quality

Optimalmaintenance

strategy

Raw data

Denoising

Classification

Feature extraction

Framework of Prognostic and Health Management(PHM)

Information fusion

4. PHM for traction substation

40

Multi-source and multi-scale data of PHM



41

Data processing and data fusion


42

Fast fault diagnosis;

Fault prognosis;

Health assessment;

Reliability analysis;

Risk assessment;

Maintenance decision

PHM algorithms and applications:



43

Fast identifying the fault components Diagnostic model- Fuzzy Petri Net (FPN) Considering the reliability of information source devices

Topology structure of traction power supply system

Research Content: Fast fault diagnosis

11w

12w

13w14w

16w

17w

1t

2t

3t

1P

2P

3P

4t

5t

8t

9t

21w

22w

23w

24w

31w

32w

10P

11P 1PT

1

2

3

4

5

6

7

4P

5P

25w

15t

16t

8

9

6t

7t

6P

7P

10t

12P

10

26w

27w

33w

8P

9P

11t

17t

12t

11

12

13

13t

15w

18w

19w

110w

111w

112w113w

114w

115w

116w

117w

118w

28w

29w

210w

211w

14t14

15

16

17

13P

14P

Fuzzy Petri Net model of traction power supply system

0 1 2 3 4 5 6 7 8 9x 104

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

t1/小时

R(t

)

t2=1

t2=45000

保护的失效函数

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5x 104

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

t2/次数

R(t

)

t1=1

t1=90000

断路器的失效函数

Reliability of protection device

Reliability of circuit breaker



44

Prognosis-predict faults in advance

Prediction method based on HMM+SVM

Research Content: Fault prognosis

Small size sampledata applicable

Prediction method-Hidden Markov Model (HMM)

+ Support Vector Machine (SVM)



45

Health value and its significance

Research Content: Health assessment

1-0.9600 0.9599-0.9 0.8999-0.8 0.7999-0.6 0.5999-0

Health Subhealth Slightly morbid Medium morbid Serious morbid

Safe Deteriorating

No faultEarly latent hazard Late latent hazard Functional fault

Fault

Health value

Health status

Service Performance

Fault evolution

Failure

Hidden fault

Health value reflect performance and fault situation ofequipment and system

Assessment model- Analytic Hierarchy Process (AHP): fromequipment layer to system layer

Algorithm- Fuzzy statistics: eliminate subjective uncertainty

Health assessment model and algorithm



46

Reliability deteriorating curve

Research Content: Reliability analysis

Reliability analysis based on equipment deteriorating

Analysis model- Markov reliability model

Considering various weather condition

Reliability analysis results

12P

11P 22P

23P

33P

34P

44P

11

xe 22

xe 33

xe 44

xe

1t 2t 3t 4t

13 fN fN

23 fN1

1t 2t 3t

(a) Ignore weather condition

Initial dataFitting curve

0

2

4

6

8

10 x104x10

5

0

2

4

6

8

0.9993805 0.9993815 0.99938250.998750 0.9987660.9987580.998742

x104

0.994340 0.994375 0.994390

2

4

6

8

0.9145 0.9147 0.9149 0.9151

1

3

5

7x10

3

Probability density

Reliability


Reliability(b) Normal weather condition

Probability density


Reliability

Probability density

(c) Bad weather condition


Reliability

Probability density

(d) Disaster weather condition

Markov Reliability analysis model



47

Traction power supply system lightning risk area illustration

Research Content: Risk assessment

Lightning risk assessment for traction substation andoverhead catenary

Direct lightning risk + Induced lightning risk

Primary equipment risk + Secondary equipment risk

变电所建筑物

H 1:3

3H

W

L

(×10-3)

0.00.51.01.52.02.53.03.54.04.55.0

Traction substation Overhead CatenaryPrimary equipmentrisk by Direct Lightning strike

Secondary equipmentrisk by Direct Lightning strike

Equipment risk by induced lightning Total lightning risk

Ligh

tnin

g ris

kva

lue

lower limit

Upper limit

Risk interval

Lightning risk assessment results



48Research Content: Maintenance decision

Maintenance strategy

Cost

Health (Reliability)

Inspection cost

Downtime lossMaintenance cost

Condition-based maintenance curvex

y

z

M

M1

M2

Health

Reliability Safety

Maintenance EconomyNo action？postponed？Prioritized？ASAP？Immediate？

Maintenance time decision

Breakdown maintenance (BM)?Preventive maintenance (PM)?Hazard detection (HD)?Condition based maintenance (CBM)?

Maintenance mode decision

Period of PMFrequency of Manual inspectionInspection frequency of potential hazardInspection frequency of hidden fault

Maintenance period decision

No maintenance?Minimum maintenance?Imperfect maintenance?Perfect maintenance?

Maintenance extent decision

Various maintenance decision parameters

Five maintenance properties

Multi-objective optimization method

Five maintenance properties

Multi-objective optimization

Example of maintenance decision parameters



49PHM and data management software

Integrated informationdatabase

Visualization module

Data analysis and process module

Maintenance decision module

Raw data

Customized data

Mutual support

Report forms

Generate & Print

Software input module

Raw dataEquipment importance evaluation module

Health assessment module

Data management module

Generate history data

Maintenance mode decision

Maintenance period decision

Maintenance optimization

Fault prognosis module

Fast fault diagnosis module

Software structure frame diagram



50

Main interface of health assessment

Software interface(1) Health assessment module

Equipment configuration information

Health assessment results

Equipment health trend

PHM and data management software4. PHM for traction substation


51

Software interface(2) Fault prognosis module

Equipment fault degradation trend

Main interface of fault prognosis Prognosis and condition monitoring results



52

Software interface(3) Fast fault diagnosis module

Fast fault diagnosis results

Main interface of fast fault diagnosis Recommended maintenance strategy



53


Conclusion

School of Electrical engineering in Southwest Jiaotong

Univ. has achieved rich results in protective control, power

supply monitoring and dispatching, pantograph-catenary

system and PHM for high-speed railways.

Simultaneously, the control theory and electronics

technologies are used in the state-of-the-art technologies in

our school, such as co-phase supply system, Maglev

system, wireless traction power system.

Thanks for any

suggestions！


Protective Control and Fault Diagnosis of High-speed Railway ...

Documents