Information Technology Reference
In-Depth Information
numerical simulation in such a large scale system becomes an intractable task.
Through model reduction to choose appropriate control structure is an approach
to overcome this problem[2].
Model reduction methods have become increasingly popular in recent years.
There are a number of mathematical formulations and systematic strategies pro-
posed for the model reduction of the HVAC system. One of the most common
model reduction schemes is balanced truncation which was first introduced by
Mullis and Roberts (1976)[3], and later Moore (1981) applied it in the systems
and control literature. Furthermore, this approach was extended by He[4] to the
controller design of vapor compression cycle. Unfortunately no validation re-
sults were presented to prove the accuracy of the reduced order model. Another
popular model reduction method is the balanced residualization introduced by
Fernando K and Nicholson H[5]. A 4th-order model for the transcritical vapor
compression system was proposed by Rasmussen[6] using residuatization princi-
ples and dimensional hankel singular value perturbation method to remove the
redundant mass balance state. The model preserved the physical meaning of the
dynamic states but still presented some complex challenges for the controller
design and implementation. Based on Laguerre polynomials, Wang[7] proposed
new methods for model reduction of coupled systems in the time domain. By
defining projection matrices according to laguerre coecients, reduced order cou-
pled systems are generated to match a desired number of these coecients. The
new methods retained the stability of coupled systems, but didn't implement on
an actual HVAC system.
In this paper, a structure selection criterion, which can be used to choose
the optimal control structure and evaluate the control performance of differ-
ent control structures, is proposed. The remainder of the paper is organized as
follows. Section 2 details the dynamic model of the vapor compression cycle sys-
tem. Section 3 proposes a structure selection criterion which is used to evaluate
the performance of different control structures and choose the optimal simplify
model. Section 4 presents the experimental results which justify some of the
conclusions discussed in the previous sections. Section 5 summarizes the main
conclusions.
2 Dynamic Modeling and Model Linearization
A typical single vapor compression cycle system is shown in Figures 1. Using the
lumped-parameter and moving-boundary method, the dynamic model of each
component of vapor compression cycle was derived by B.P. Rasmussen and is
brief listed below[6].
1) Compressor: The dynamics of the compressor is considered to be much
faster than those of heat exchangers, therefore, its mass flow rate can be modeled
as a static component
˙
m
k
=
ω
k
V
k
ρ
k
1+
C
k
+
D
k
P
ko
P
ki
(1)
Search WWH ::
Custom Search