Sumários
Sensitivities, radial networks, probabilistic power flow
26 outubro 2017, 11:00 • Luis Marcelino Ferreira
Review of loss sensitivities
Example of a situation in which one could compute sensitivity of losses wrt power injections
Radial networks and distribution system
Why a radial structure for networks
Trees and the relationship parent-child
Power flow for distribution networks
Upward sweeping for currents
Downward sweeping for voltages
Convergence criterion
Loads are uncertain: fuzzy vs probabilistic
Radially currents are random variables and what adds up are random variables
Types of random variables for loads
Mean values add up, as E(.) is a linear operator
std(.) is not linear
Near the top (of the tree), the current’s distribution is more concentrated near the mean value: smaller std
For security reasons the system is planned for values far from mean values (2 std’s); numerical values for currents in a node do not add up
Probabilistic analysis if the network is meshed: an introduction to Monte Carlo
What are the final results of a probabilistic analysis
n>Contingency analysis; losses
24 outubro 2017, 09:30 • Luis Marcelino Ferreira
Revisions
N-1 criterion for system security
Single contingency
MIL and factor updating
Compensation
More details on compensation methods
Review of its circuit principle
Line powers
Fictitious injected powers
How to find the error
How to set up the equation to correct for error
Problem dimension
Newton’s numerical application
Comparisons to make the principle easily acceptable
Another subject: Losses
The role of losses in engineering planning
How to assess losses
How to compute losses traditionally (B method)
How to compute losses today
How to compute derivatives for losses (or sensitivities): traditionally and by using power flow results and power flow solution, including pf JacobianHow to compute those derivatives using a power flow program and the definition of derivative