Page 232 - Mathematical Models and Algorithms for Power System Optimization

224 Chapter 7

The power system is a complex large system; some literatures have considered the multiunit
system but only employed a method of centralized state feedback control. Such a
method needs an appropriate amount of information, without which it is hard to put into
practice application.

In addition, the existing literatures only made the load step disturbance simulation test.
As previously mentioned, the power system operation process is a random process, and
the load disturbance is a process of continuous change. The system state changing
under such a disturbance is also a continuous change process. Therefore, it is not
enough to only study the system changes under the step disturbance, because the resulting
control law and control method may not be able to meet the requirements for actual
control performance.

7.1.2 Overview of this Chapter

Proceeding from the stochastic control theory, based on the applications of the identification,
estimation, and control theories, and combines computer technology, this chapter studied the
load frequency control problem in the normal operation of the power system. As previously
mentioned, the load disturbance of the power system is a complex type of disturbance. First, this
chapter studies the power system load disturbance model identification problem. A time series
correlation function analysis method was applied to determine the model structure and its
parameters of the load disturbance, and a mathematical model of the load disturbance was
established.

Second, due to the complexity of the power system, this chapter adopts the idea
of decomposition and stratification to construct the state estimators at different levels, so as
to estimate the operating status of the system and predict the trend of the load disturbance.
Finally, the law of load frequency control is derived by applying the principle of invariance,
and a load frequency compensator (i.e., feedforward controller) is designed. Based on
the feedforward controller, a digital computing simulation of the load frequency random
tracking control is performed to implement system frequency control for an actual power
system structure.

The performance of the controller designed by this principle is much better than that of
the integral feedback controller. Moreover, this chapter also handles all linear model
transformations required for identification, estimation, and control. The transformations
include those from difference equation to differential equation, and from differential
transfer function to difference transfer function. Some of these problems have not been
proposed or yet solved before, and the solution of these problems will bring us great
convenience.