Page 319 - Decision Making Applications in Modern Power Systems

280 Decision Making Applications in Modern Power Systems

determined by the reserve market. These different markets can either be
obtained sequentially in the unbundled market systems or in the same opti-
mization problem in the integrated market systems [32].InRef. [33] the
effectiveness and the advantages of both systems are assessed. In reality the
sequential approach is usually applied. However, this approach gives a sub-
optimal optimization solution to the overall objective, and feasibility issues
can also arise. For example, if the reserve schedule is first determined while
neglecting the N 2 1 criterion and all the reserves are allocated to the cheap-
est generator, therefore there is no feasible solution to an N 2 1 secure
energy scheduling if this generator is tripped, since no other unit can provide
the reserves that are required to compensate its production.
This is one of the worst-case scenarios, which shows that the reserves
may not be adequate in unbundled market systems. However, in practice,
heuristics are used to take care of such extreme issues. As a result, an inte-
grated market mechanism allows us to ascertain the optimal solution to the
overall problem. In this line a framework dealing with the cooptimization of
energy and reserves, which takes into account network constraints and the
N 2 1 security criterion is developed. In Refs. [32 37] the reserve optimiza-
tion for a security-constrained market clearing context while maximizing the
expected social welfare is presented. In Ref. [32] a multistage stochastic
unit-commitment program, which models the uncertainty in generation by
using reduction techniques to ensure tractability of the problem, is outlined.
The limitation of these methods is that they do not guarantee reliability of
the resulting solution.
This section presents a unified framework that simultaneously solves the
problem of designing an N 2 1 secure day-ahead dispatch for the generating
units, while determining the minimum cost reserves and the optimal way to
deploy them. A probabilistic methodology which guarantees the satisfaction
of the system constraints is used to account for wind power inconsistency.
The security constraints emanating from the N 2 1 criterion are first inte-
grated to a DC-OPF problem and formulate a stochastic optimization prob-
lem with chance constraints.
By modeling the steady-state behavior of the secondary frequency con-
troller, LFC controllers, the reserves can be represented as a linear function
of the total generation-load mismatch. Generation-load imbalance can be a
result of difference between the actual wind and its forecast, or a generator
load loss.
Different ways of reserve distribution, which are based on the type of
mismatch offering an implementation of corrective security, are introduced
in literature. The overall objective formulation includes both preventive and
corrective control [33]. Preventive control actions are the generation dispatch
and the reserve capacity determination, while the contingency-dependent
reserves allocation in real-time operation is corrective control. The advan-
tages of these strategies are their physical intuition and the decision