Framework for comparing power system reliability criteria

Report
Framework for comparing power
system reliability criteria
Evelyn Heylen
Prof. Geert Deconinck
Prof. Dirk Van Hertem
Durham Risk and Reliability modelling for Energy
Systems day
November 12th, 2014
Introduction
Current situation in power system reliability management
• Deterministic N-1 criterion with various shortcomings
• Major evolutions in the power system
• Increasing uncertainties
Probabilistic reliability management
• Take into account probabilities
• Could tackle shortcomings of N-1
• Many academic references
• Not fully used in practice
 Amongst others due to lack of quantified benefits
Framework for
Comparing power system reliability criteria
2
Outline
• Framework for comparing power system reliability criteria
o
o
o
o
o
Overview
General schematic of the implementation
Implementation of deterministic reliability management module
Assumptions
Comparison of deterministic criteria
• Conclusion
3
Outline
• Framework for comparing power system reliability criteria
o
o
o
o
o
Overview
General schematic of the implementation
Implementation of deterministic reliability management module
Assumptions
Comparison of deterministic criteria
• Conclusion
4
Framework for comparing reliability criteria
Objectives of the framework:
1. Quantification of performance of various power system reliability
criteria and their management
2. Comparison of performance
3. Identifying alternative reliability criteria
5
General schematic
6
Deterministic reliability management module
1. N-1 criterion: ‘System
should be able to withstand
at all times the loss of any
one of its main elements
(lines, transformers,
generators, etc.) without
significant degradation of
service quality.’
1
2
2. State enumeration:
• Run power flow
• Check for operational
limit violations
3. Decision = balance
reliability and cost
4. Reliability actions
• Corrective actions
• Preventive actions
7
4
3
Reliability management modules
Reliability
assessment
methods
Reliability
criteria
Reliability
control
Analytical contingency
enumeration
N-k
Preventive/corrective
control
Event tree/fault tree
analysis
Optimization
Random sampling
(Monte Carlo)
Markov analysis
8
Limits on reliability
indicators
Optimization and limits
on reliability indicators
Asset management
System development
Reliability management modules
Reliability
assessment
methods
Reliability
criteria
Reliability
control
Analytical contingency
enumeration
N-k (i.e. N-0 and N-1)
Preventive/corrective
control
Event tree/fault tree
analysis
Optimization
Random sampling
(Monte Carlo)
Markov analysis
9
Limits on reliability
indicators
Optimization and limits
on reliability indicators
Asset management
System development
General schematic
10
Framework for comparing reliability criteria
11
Data generation
• Unit commitment model
• Monte Carlo
Data modules
• Matlab m-file
Events and triggers
• Input reliability assessment
12
PF = Power flow
SW = Social welfare
OPF = Optimal power flow
Test Setup (I)
• Reliability criterion e.g. N-1, N-0
 Extended problem formulation using
islanded systems
• Reliability control
e.g. preventive, corrective
 Interlinking constraints between
islanded systems
13
PF = Power flow
SW = Social welfare
OPF = Optimal power flow
Test setup (II)
• Objective function
 Minimal cost for society == maximal
social welfare
• Probabilistic approach in large
system
 State selection
14
PF = Power flow
SW = Social welfare
OPF = Optimal power flow
Simulation
• Optimization, e.g. SCOPF, OPF
 Economic dispatch of generators
satisfying operational limits, reliability
criterion and control constraints
15
PF = Power flow
SW = Social welfare
OPF = Optimal power flow
TSO actions
• Short term: Preventive and
corrective actions
• Medium term: asset
management & operational
planning
• Long term: system development
16
Reliability assessment
• Check performance of reliability
criterion and reliability control
using PF and OPF for:
• All contingency cases
• Contingency cases of truncated
state space
• Specific scenarios (i.e. events)
17
Events
• Evaluate performance of reliability management for specific
cases
• Time series including results of events due to:
o
o
o
Natural hazard
Operational conditions
Human behaviour
• Which can lead to:
o
o
o
o
18
Discrepancy between generation and load
Generator/branch outage at particular moment in time
Failure caused by several simultaneous faults (failure of cable or
power line in same trace etc.)
…
Economic evaluation
• Based on market model
• Social welfare evaluation
• Total cost evaluation
 Could be substituted by more
complex evaluation techniques
19
Comparison of reliability criteria
• Methodology for comparing reliability criteria
• Appropriate metric for comparing reliability
criteria
20
Assumptions
•
•
•
•
•
•
•
•
•
No generator ramp rates or minimal on/off times
Knowledge of Value of Lost Load at every node
Linear cost curves  constant marginal costs of different units
Corrective actions
o Generation redispatch
o Load shedding
Constant failure and repair rates  exponential distribution
Aggregated branch models
No failure of corrective actions
Reliability assessment considers only branch outages
No forecasts errors included (wind, load…)  No stochastic, multi-stage
optimization
• Single TSO, Single area
• DC power flow
21
Comparison: Results
• Three node test system
• Comparison of
N-0 corrective
o N-1 preventive
o N-1 corrective
Varying value of lost load (VoLL)
o
•
Performance of reliability criteria
and their management
dependent on VoLL
22
Outline
• Framework for comparing power system reliability criteria
o
o
o
o
o
Overview
General schematic of the implementation
Implementation of deterministic reliability management module
Assumptions
Comparison of deterministic criteria
• Conclusion
23
Conclusion
• Comparing power system reliability criteria is important
• Framework for comparing power system reliability criteria
and reliability management
o Objectives of the framework:
1. Quantification of performance of various reliability
criteria and their management
2. Comparison of the performance
3. Identifying alternative reliability criteria
o Quite complex, even with many assumptions included
• Preliminary result: Performance of reliability criteria and their
management dependent on VoLL
24
Thank you!
Questions?
[email protected]
[email protected]
[email protected]
The work of Evelyn Heylen
is funded by:
Research in the framework
of the Garpur project

similar documents