Improving Special Protection Schemes with Operations ...

13/11/2017 1

Improving Special Protection Schemes

with Operations Research analysis

Code: 02.047

Alejandro Hoese

EEConsulting, San Juan, Argentina

13/11/2017 2

1. What are Special Protection Schemes (SPS) ?

2. SPS Main Characteristics, Design and Challenges

3. Operations Research (OR) Techniques

4. Two Examples

5. Conclusions

Contents

13/11/2017 3

- Automated System

- Detects abnormal conditions (system studies)

- system instability;

- unacceptable equipment tripping;

- violation of voltage levels or loadings on transmission facilities;

- unacceptable loss of load or generation.

- Acts along with ‘conventional’ protection systems

- UFLS (underfreq load shedding) and/or UVLS;

- Protective relays;

- AGC (automatic generation control), etc.

- Takes Preventive/Corrective Actions

- system reconfiguration (line/gen/load tripping);

- generation runback;

- reactive power, etc.

- Helps delaying expensive transmission infrastructure

Special Protection Schemes (SPS)

13/11/2017 4




4. Two Examples

5. Conclusions

Contents

13/11/2017 5

- Control actions faster than human interventions;

- ‘Ad hoc’ solution roughly applied to another power system;

- ‘All time’ operative normal, alert and emergency op. cond.;

- Automatically armed after system restoration;

- Malfunctions should not lead to higher costs !!

- PS growing complexity SPS more sophisticated

- There is no unified definition:

- SPS = Remedial Action Scheme (NERC)

- SPS = WAPS (WECC) control of an extensive area

- SPS AGC / ALS / SFC

- SPS protection reliability security

SPS Main Characteristics

13/11/2017 6

- Traditional design approach:

- Define possible and probable contingency scenarios

- System Analysis security limits for relevant states

- Build predefined ‘tables’: scenario action

- Traditional behavior:

- Identify predefined scenarios real-time assessment

- React according predefined actions

- Finite relevant states define ‘ranges’ of variables

- Temperature, Ampacity, Frequency, Voltage, Topology…

Suboptimal operation

Op.states increase exponentially with growing areas

SPS Design and Challenges

13/11/2017 7

- Real-time assessment (NERC)

- existing (pre-contingency) and potential (post-contingency)

operating conditions

- ¿Can be really defined

- EVERY possible scenario ?

- EVERY corresponding control action?

- New design approach:

Intelligent real-time assessment math model of the network

1. Pre-Calculate specific remedial actions for predicted

a) possible contingencies for current scenario;

b) post-contingency (controlled) oper. conditions;

2. After contingency: a) rearm if success; b) disable otherwise;

c) enable for new stable or controllable operating conditions.

SPS Challenges

13/11/2017 8

Traditional Real-Time Assessment

POWER SYSTEM + WEATHER COND.

Communication Network

RTAC

RTUs

Elec. Studies

Off-Line

Scenario/Control

13/11/2017 9

Intelligent Real-Time Assessment

POWER SYSTEM + WEATHER COND.

Network Models

Decision Making Models

Communication Network

RTAC

RTUs

On-Line

Elec. Studies

Off-Line

< 800 ms

13/11/2017 10




4. Two Examples

5. Conclusions

Contents

13/11/2017 11

SPS fast response (< 800 ms)

wide area networks high latency

asynchronous data

network models non-linear

Control Actions Best you can get with real restrictions

Find best suboptimal solution

OR Techniques

OperationsResearch

Optim. Algorithms

Optim. Heuristics

Statistics, Estimators…

Probability, Fuzzy…

Real Networks

13/11/2017 12

Complex SPS Engineers

Mathematicians + Programmers

IT + Communication Technicians

RTAC high computational power

lack of development tools

Traditional approaches same response for similar inputs

easy programming and testing

OR approaches complex software and responses

hard programming language (IEC 61130)

hard debugging, testing and tuning

tuning longer dry-run operation

OR Techniques

13/11/2017 13




4. Two Examples

5. Conclusions

Contents

13/11/2017 14

Example 1: Congestion Management

T

Paposo

Maitencillo Nogales

550 km 500 km

> 1.5 GW Renewables

Control Node

220 kV

13/11/2017 15

- High Latency asynchronous measurements

- > 30 generation plants (wind + solar)

- Nine 220kV SSEE (power flows, topology, temperature)

- Control Actions

- Generation Control maximize transmission capacity

control ampacity limits

- Generation Disconnection contingencies (overloads)

< 800 ms (protection relays)

- Different network topologies post-contigency control

network expansion

- Automatic arming, rearming and disabling:

disable under measurement inconsistency

(failures, transients, etc.)

temporarily disable after contingency control

enable under stationary conditions


13/11/2017 16

- Network model: DC with active losses

- Uncongested network: increase/decrease set points for every Gen

maximize use of lines

MIP non-recursive optimization heuristic

- Congested network: new setpoints for every Gen (heuristic)

ampacity model

ambient temperature (statistical back up)

solar condition (astronomic model)

- Overloads: contingencies out of the supervised network:

high overloads disconnect min. generation (comb.)

low overloads change setpoints (heuristic)

- Overloads: contingencies inside the supervised network

disconnect min. generation (comb.)

< 800 ms (before action of protection relays)


13/11/2017 17

Example 2: Increase transmission capacity

2x220kV23 km

Weather Station

SE Ancoa 220kV

Run-of-the-river generation

HD Pehuenche

J4 J5

J1 J2

JU1 JU2 JS

SE Pehuenche

SIC

T1 T2

OPLS1 OPLS2

Unwired Input Wired Input Wired Output Wired I/O

13/11/2017 18

- Maximize transfers Operation behind N-1 criteria

- Real-time thermal rating (IEEE 738-2012)

- Control of power flows through Pehuenche Hidro

2 x 550 MVA units

Normal operation: 140 MW per min (both units)

SFC operation: 0 MW per min (control unit)

50 MW per min (other unit)

- Response time: function of current & temperature (IEEE 738)

- Very high reliability redundant measurements:

5 dynamic state estimators + 1 statistical model

i) ambient temperature Ta [ºC] – Statistical Model

ii) solar global radiation R [W/m2] – Dynamic Least-Squares (DLS)

iii) conductor current [A] Ic(Tc) – DLS model (Fourier);

iv) conductor temperature [ºC] Tc(Ic) – DLS model (Fourier);

v) thermal emissivity coefficient estimation (Markov chains);

vi) solar absortivity coefficient estimation (Markov chains)

Example 2: Increase transmission capacity

13/11/2017 19




4. Two Examples

5. Conclusions

Contents

13/11/2017 20

1. Modern SPS in transmission:

- Push equipments to operative limits

- High reliability every possible scenario

different topologies

measurement bugs

- Complex tasks optimizations OR techniques

statistical analysis

pattern recognition

detailed network modeling…

2. Design & Development Mathematics + Engineering

3. Available Hardware (RTACs) and Communication networks

4. Lack of development tools (libraries, languages, debugging

environments)

5. Virtual controllers behave different as real ones

6. Longer times for SAT and supervised operation (dry-run)

7. Detailed back-up of events recommended

Conclusions

13/11/2017 21

Thank you!

Improving Special Protection Schemes with Operations ...

Documents

Transcript of Improving Special Protection Schemes with Operations ...