You are on page 1of 12

Building and Environment 127 (2018) 1–12

Contents lists available at ScienceDirect

Building and Environment


journal homepage: www.elsevier.com/locate/buildenv

A new simplified modeling method for model predictive control in a T


medium-sized commercial building: A case study
Junhua Zhuanga,c,∗, Yimin Chenb,c, Xiangguang Chena
a
School of Chemistry and Chemical Engineering, Beijing Institute of Technology, Beijing, China
b
Department of Civil Architectural and Environmental Engineering, Drexel University, Philadelphia, PA 19104, USA
c
School of Electrical and Information Engineering, Beijing University of Civil Engineering and Architecture, Beijing, China

A R T I C L E I N F O A B S T R A C T

Keywords: Model predictive control (MPC) methods for heating, ventilation and air conditioning (HVAC) systems have been
HVAC studied to improve the control accuracy and reduce energy consumption in recent years. The accuracy of the
Power spectral density model for building thermal dynamics in MPC plays a critical role to accurately control the system. The modeling
Model predictive control method also impacts on the real practice of MPC in buildings due to its cost and scalability. Studies have shown
Feedforward control structure
that an appropriate simplification of modeling procedure has minor impacts on the model accuracy, but in-
creases the modeling efficiency. In this article, variables including weather conditions, occupancy and electricity
are divided into two categories: manipulated variables and random variables. A novel two-step modeling
strategy is proposed for simplifying modeling procedure and increasing model accuracy. Manipulated variables
are used in step response method to develop system model. A low order system is obtained after the model
simplification by observing the response curve. Random variables are used in the power spectral density (PSD)
method for modeling. Transfer function is obtained through calculating the cross-power spectral density (CPSD)
of the system output and input, the PSD of the input, and the ratio of CPSD and PSD. A MPC strategy with
feedforward control structure is proposed to utilize the obtained dynamic characteristics of random variables
and effectively compensate the errors caused by these variables. Field test in a medium-sized commercial
building is implemented to evaluate the MPC strategy. The result shows that a considerable amount of energy
saving is achieved through the proposed MPC.

1. Introduction HVAC optimal control [3]. Khanmirza et al., presented a comparison of


different control method for a building hybrid heating system, and
In China, buildings are responsible for 28% of the total energy proved that maximum cost reduction can be achieved while indoor
consumption [1]. In commercial buildings, heating, ventilation, and air temperature requirement can be met by implementing the MPC method
conditioning (HVAC) system accounts for more than 60% building with economical optimization [4]. Salakij et al., developed a MPC
electricity consumption to meet the demand of occupants' comfort. method to incorporate critical building information into control algo-
Therefore, a potential benefit can be achieved and environmental im- rithms based on a lower order system model modified from coupled
pacts can be reduced through energy saving in building HVAC systems. system model [5,6]. Through simulation and experiment validation,
Compared to HVAC equipment update in existing buildings, the im- they demonstrated that the proposed MPC method yielded superior
provement of HVAC control strategies is considered as one of the effi- control performance with lowest energy consumption while main-
cient and cost effective ways to increase the building energy efficiency taining indoor thermal comfort.
[2]. Although there are extensive researches so far on the development
Although traditional control methods which include the propor- of MPC for increasing energy performance in buildings, there are only a
tional-integral-derivative (PID) control and the ON/OFF control are still few real practice examples have been reported from literature. Huang
the most commonly used control methods in most commercial build- et al., implemented MPC in an airport terminal building. The simulation
ings, these methods have drawbacks in that system level characteristics and field experiment demonstrate that energy saving can be achieved
and multiple interactions among different components are not con- without losing thermal comfort after using the proposed hybrid model
sidered. In recent years, MPC has witnessed a rapid development in the predictive control [7]. Ruano et al., introduced the implementation of


Corresponding author. School of Chemistry and Chemical Engineering, Beijing Institute of Technology, Beijing, China.
E-mail address: zhuangjh@bucea.edu.cn (J. Zhuang).

http://dx.doi.org/10.1016/j.buildenv.2017.10.022
Received 2 July 2017; Received in revised form 27 September 2017; Accepted 16 October 2017
Available online 18 October 2017
0360-1323/ © 2017 Elsevier Ltd. All rights reserved.
J. Zhuang et al. Building and Environment 127 (2018) 1–12

an intelligent model based predictive control (IMBPC) system which state-space which results in great difficulties in practical application
consists of the software and hardware in a real university campus [43]. For the purpose of chiller plant control, it is not necessary to
building [8]. Through the experiment for almost one month, HVAC model all zones in the building. Moreover, the structures and para-
electrical saving can be reached and thermal comfort is also maintained meters of a RC model obtained from one building can hardly be em-
by using the IMBPC. ployed in other building. This makes the scalability of a RC model to be
MPC achieves higher control quality compared with PID or ON/OFF lower in the real practice [11]. Therefore, some simplified methods
control due to its three features including: predictive model, rolling have been proposed to cope with this issue. For example, Braun et al.,
optimization and feedback correction [9,10]. The procedure of mod- modeled a whole large sized building as a single zone instead of de-
eling is a critical prerequisite for a successful implementation of MPC in veloping a multi-zone RC model [25]. The result indicated that there
buildings [11], because a control process will react to the prediction of was a close match between the predicted cooling load and the actual
input changes or other disturbances [12]. Modeling can be divided into cooling load. Sourbron et al., also demonstrated that the second-order
physical-based modeling (or white box modeling) method and data- model can achieve equal control performance as the forth-order one
driven based modeling (or black box modeling) method. Recently, data- [20]. Richard et al. proved that it is feasible to simplify high order
driven modeling methods have been widely utilized to model HVAC systems to low order ones in engineering applications [12,27,28].
systems for their advantages on cost-effectiveness and flexibility [13]. The system parameters identification (SPID) is another challenge
Artificial neural network (ANN) is one of widely employed data-driven after the determination of system structure. Various SPID methods have
method in building energy prediction (modeling) due to the ability to been proposed to optimize model parameters. O'Dwyer et al., developed
deal with nonlinear, multivariable modeling problems. Afram et al., a system identification method from data in which significant un-
gave a good review on artificial neural network (ANN) based MPC measured disturbances are present [29]. In this method, high-order
system design [14]. A dynamic temperature setpoints profiles of the simulation models for control strategy analysis and low-order zone
zone air and buffer tank can be generated through MPC to save oper- models for optimization were developed separately for MPC. Karlsson
ating cost. Huang et al., developed an ANN model-based system iden- et al., identified the parameters by using step-response method and
tification method to model multi-zone buildings [15]. The model in- pulse-response method [22]. It is effective to model the indoor tem-
cludes different energy inputs to increase model accuracy for MPC. perature influenced by the supply heat flux, since the supply heat flux
Ferreira et al., proposed a discrete model-based control method which can be a step change or an approximate ideal impulse function. How-
include three components as predictive models based on ANN, the cost ever, these methods that are called experimental methods can only be
function aiming to minimize energy consumption and maintain thermal applied to variables that can be manipulated (e.g. the supply heat flux,
comfort and a discrete branch and bound approach for optimization the supply chilled water temperature), and can't applied to random
[16]. ones (e.g. the outdoor temperature, the solar radiation). By using si-
As for physical-based method, it can provide detailed building in- mulation, Fraisse et al., studied the internal surface temperature re-
formation based on the analysis of building material features. So some sponse after a step change in external or internal temperature [30]. Yao
researchers proposed that rigorous model or physical model should be et al., also studied the dynamic response of air temperature in different
implemented [17–19], but it is seldom employed in real engineering zones when subject to a step change of ambient temperature, occupant
practice for the high costs and expert involvement. Furthermore some number and etc. [31]. However, some weather variables such as out-
key parameters such as the heat transfer coefficient and specific heat door temperature, solar radiation, cannot be step changed in real
capacity of building envelope cannot be easily obtained, especially in practice.
large size commercial buildings, and when there exists high couples Data-driven methods for model parameters identification can be
among different components in a system. Fortunately, some studies applied after system structures are determined (e.g. a simplified lower-
show that a simplified model can achieve acceptable accuracy in order models). These data-driven methods are better at handing
comparison to a detailed physical model for control [20–23]. Sourbron random variables such as weather prediction data. Various methods
et al., established a second-order model and a fourth-order model, including least squares method [32], regression method [33], Kalman
which included additional solar and internal gains in the identification filter [23], ensemble Kalman filter [34], ARMAX model identification
data set to improve performance as contrast to the second-order one, for and subspace identification [35] have been employed to estimate the
concrete core activation (CCA) [20]. The result indicated that both model parameters. Models developed through random variables can be
second-order and forth-order model can achieve a satisfactory control obtained by these methods. Some research works treat manipulated
performance. Bălan et al., demonstrated a simple solution with a variables such as temperatures and flow rates of supply chilled water, as
second-order model for thermal modeling of a house which included random variables [13]. For example, Li et al., pointed out that model
experimental identification of the model's parameters. Based the model, accuracy and calculation efficacy cannot be guaranteed through the
they successfully used a predictive control algorithm to control the model identification methods which lack of active building excitation
thermal system in a house [21]. Karlsson et al., established a simplified [13]. This is because database data does not contain plenty of useful
two-node lumped model for a single room [22]. Prívara et al., identified information about systems, in other words, data rich and information
a real building model by using subspace methods and successfully ap- poor [36]. Fig. 1 shows a summary of models and parameter identifi-
plied it into real control [23]. cation methods mentioned above.
Similar to these low-order models stated above, resistance and ca- Although a number of successful cases of modeling can be found,
pacitance (RC) network models or inverse models are widely used for some issues still need to be solved to apply MPC in the real practice.
building energy estimation [24,25]. RC network modeling possesses a These include:
meaningful and transparent configuration, making the developed
model more understandable and reliable to use [7]. Since the models • Models inevitably have errors. The system control error cannot be
established by the RC method or some simplified methods are all linear eliminated successfully even with high-order models because it is
models, it is actually already using a priori knowledge – the system can difficult to determine the model parameters.
be approximated as a linear system, when using these methods. As for • The presence of external uncertain disturbance such as weather
large sized buildings, state-space modeling approach based on RC prediction has significant impacts on the control quality.
method can also be applied [26], although sometime this approach are
applied in a single zone building model for details [25]. In practice, a Concerning the above problems, this article aims to develop a
large sized building includes many zones, rather than two zones as in strategy of MPC with feedforward control structure, which has both the
Ref. [26]. Many zones will produce a high dimension matrix of the MPC feedback structure and the feedforward structure [37,38]. As

2
J. Zhuang et al. Building and Environment 127 (2018) 1–12

Fig. 1. The summary of models and parameter identification.

mentioned above, the inputs of the controlled system can usually be


divided into two categories, one is manipulated variables, and the other
is uncontrollable variables, which are often random variables. The
purpose of control is to continuously adjust manipulated inputs to
overcome the influence of uncontrollable inputs, and make the output Fig. 2. The working flow of this study.
has the desired dynamic characteristics. For irregular disturbances such
as measurement errors, this adjustment can only be made if it is re- manipulated variables, where the system model is the bypass valve
flected to the detectable variables such as outputs, and therefore the control model. PSD method is applied to model the system under
feedback method must be used. However, for those random variables random variables that are outdoor temperature and solar radiation. In
(e.g. weather condition) that can be predicted, it is unreasonable if they applying the PSD method, some parameters may be adjusted to make
can be corrected by feedback as disturbances. The feedback control the model more matched with the real system. These models are ex-
only works when the error has impact on the output, which results in a pressed in the form of transfer functions, and then applied in a MPC
certain hysteresis. Furthermore, the feedback control does not take full method with feedforward control structure to the system. The whole
advantage of the known information of this part of the inputs. Thus, it is flow of the study is shown in Fig. 2.
more reasonable to implement a feedforward control which can take
advantage of the known dynamic characteristics of this part of the in- 2.2. Building description
puts, and can effectively compensate the errors caused by these inputs
[39]. The test building is a medium-sized commercial building of a
This paper is as follows: Section 2 introduces the workflow of this shopping mall in Sichuan province, China, located at 29°35′ north la-
study including building and HVAC description. The bypass valve titude, 103°43′ east longitude. The building has 31.4 m in height and
control model is presented in Section 3, which use step response 41,000 m2 with six-floor on the ground and two-floor basement. From
method for manipulated variables and PSD method for random vari- outside layer to inside layer, the exposed walls are all made of granite,
ables. Section 4 introduces the proposed MPC strategy with feedforward thermal insulation materials, gravel concrete, hollow brick, gravel
control structure. In Section 5, a case study is used to demonstrate the concrete, and wall tile or lime. The exposed windows are all double-
effectiveness of the proposed modeling and control strategy. Conclusion glazing windows. Fig. 3 shows the external shape of the test building.
and a future research plan are given in Section 6. Fig. 4 shows the building view in Google SketchUp™, in which the di-
rection of the building can be clearly seen.
2. System description
2.3. HVAC system description
2.1. Description of working flow
The chiller plant has three centrifugal chillers with variable speed
Various inputs, such as occupancy, electricity, outdoor temperature drivers for six floors above the ground and two screw chillers for two
etc., to the HVAC system have tremendous impacts on the system be- floors underground. In this study, only the six floors part above the
havior. Various inputs should be analyzed respectively due to their ground will be studied, and the part is served by 28 air handling units
different roles and impacts. And they can be divided into some cate- (AHUs) and 22 fan coils (FCs). Each AHU has a controller that controls a
gories: constants, ignored variables, manipulated variables and random valve to change chilled water flow according to return air temperature
variables. by using the PID control method. The return air temperature is set to
In this study, two types of variable are pre-defined such that ma- 26 °C. Each FC controls chilled water flow by using ON/OFF control
nipulated variables and random variables to develop a simplified con- method according to the zone temperature directly. This study does not
trol model. The step response method is used to model the system under focus on the control of zone temperatures, for the control performance

3
J. Zhuang et al. Building and Environment 127 (2018) 1–12

Fig. 3. The test building view.

of AHUs and FCs, by and large, is perfect, which means that zone
temperatures are controlled at about 26 °C. Although, the flow through
a single FC is at an ON/OFF status which can be simplified as a discrete
value, continuous value can also be obtained for the flow through the
whole system due to the flow through 28 AHUs, that is continuous
value, is much larger than 22 FCs and 22 FCs run at different time. The
cooling rated capacity of each of the three centrifugal chillers is
2286 kW, and the consumption power of each is 437 kW. The chillers
are parallel connected. Every day, the HVAC system works from 9:20 to
21:00. Usually there are two chillers running in July and August, and
the rest of a chiller is for redundancy and maintenance in most time,
unless in the very hot weather. The plant is shown in Fig. 5. Primary
pump system is used for chiller plant which includes constant speed Fig. 5. The composition of the plant.
pumps. A bypass valve is used to balance the cooling capacity of the
plant to the internal cooling load. The bypass valve is adjusted ac-
cording to the chilled water differential pressure which is set at
0.12 MPa. The system was designed to be operated as 7 °C for the
supply chilled water and 5 °C for the temperature difference between
supply and return water.

Fig. 4. The test building view in Google SketchUp™.

4
J. Zhuang et al. Building and Environment 127 (2018) 1–12

3. Modeling procedure constant, and was not an input variable. Such a simplification does not
have much effect on the control. Similarly, the electricity power density
3.1. Determination of model inputs and outputs is also an approximately constant (value is 30 W/m2), not a variable,
since the electricity power schedule is also relative fixed.
Model input and output variable selection is crucial to the control
system. Variable selection must be consistent with the process re- 3.2. Transfer function
quirement. The accuracy of system identification and model scalability
should also be considered. Most of researchers treat zone temperatures A single-input single-output linear system can be described by a
or required amount of cooling or heat as controlled variables that is transfer function G(s) as Eq. (1) [31].
system outputs.
Y (s ) L (y (t ))
This study focuses on how to generate and deliver cooling capacity G (s ) = =
more efficiently. There is no potential for energy conservation in the U (s ) L (u (t )) (1)
chilled water delivery on the whole because the chilled water pumps Where s is the complex frequency variable that can be represented as a
are constant speed pumps. So the goal of saving energy in the primary a + bj form, Y(S) and U(S) are the Laplace transforms of the output y(t)
cooling system becomes a goal to save chiller energy consumption. The and the input u(t), respectively, and L (•) denotes the Laplace transform.
load is determined by weather, occupant and electricity, the cooling The Laplace transform is a mathematical tool used to solve linear or-
water system is not adjusted in this study, and the chilled water flow is dinary differential equations.
also approximately constant. In order to improve the efficiency of the A linear model has the properties of superposition and scalability,
chillers, there is only one way that is to increase the supply chilled which imply that the result of multiple inputs is the sum of their in-
water temperature. dividual effects. In this study, the bypass valve control model includes
The indoor temperature Ti is influenced by heat gain from outdoor three inputs and one output, and their relationship can be described as
environment Q̇out , the solar radiation Q̇sol , the internal gains by occu- Eq. (2)
pancy and electricity Q̇int , and the cooling capacity from the chiller
plant Q̇cool , as shown in Fig. 6. Controllers serving the AHUs and FCs Y (s ) = Gsup (s ) Usup (s ) + Gout (s ) Uout (s ) + Gsol (s ) Usol (s ) (2)
control the chilled water flow according to the indoor temperature, and
where Usup(s), Uout(s) and Usol(s) denote the supply chilled water tem-
then the bypass valve opening level is controlled by another controller
perature, outdoor temperature and solar radiation, respectively, Y(s) is
that serves the bypass valve for maintaining the chilled water differ-
the system output – the bypass valve opening level, and Gsup(s), Gout(s),
ential pressure. These two types of controllers are considered to be
Gsol(s) are three transfer functions of one output corresponding to three
perfect in this study. Under the condition that the supply chilled water
inputs respectively. Eq. (2) can also be written in vector form as follows:
temperature remains approximately unchanged, the cooling capacity
changes in different bypass valve opening level. A closed-loop control
⎡Usup (s ) ⎤
system is demonstrated in Fig. 6. Before the study in this paper was Y (s ) = [Gsup (s ) Gout (s ) Gsol (s )] ⎢ Uout (s ) ⎥ = G(s)U(s)
implemented, the bypass valve opening level was controlled to meet the ⎢ ⎥
building cooling load at a given supply chilled water temperature ⎣ Usol (s ) ⎦ (3)
(7 °C). But when the supply chilled water temperature increases to And the objective of modeling in this study is to find the correct transfer
improve the chiller efficiency and save energy, the bypass valve functions.
opening level decreases correspondingly. The control objective is As mentioned above, two of the three inputs are random variables
therefore set to decrease the bypass valve opening level as small as that will be modeled by using PSD method, and the remaining one is
possible to meet the requirement of internal cooling load. So the bypass manipulated variable that will modeled by step response analysis
valve opening level is used as the model output in this study. method.
As for input variables, they can be divided into two categories. One
category is manipulated variable that is the supply chilled water tem- 3.3. Model order
perature in this study. Another category are random variables that af-
fect the load, such as occupancy, electricity, outdoor temperature, solar As mentioned above, the system is a three-inputs one-output system,
radiation, etc., which are treated as input variables during modeling, and the system model G(S) includes three transfer functions Gsup(s),
and are treated as interference in control. Ref.21 points out that too Gout(s) and Gsol(s) respectively. The order of each transfer function is
many inputs cannot improve the accuracy, but increase the difficulty of determined by the tradeoff between accuracy and complexity of mod-
identification. In this study, two weather inputs – outdoor temperature eling. The increase of order leads to the improvement of model accu-
and solar radiation, which are usually considered to have the greatest racy, but increase of the complexity in modeling and control. The order
impact on indoor temperature, are selected. The occupancy schedule is of the system model is the sum of the orders of the three transfer
relative fixed, and the occupancy density is about 0.03 Person/m2 in the functions, and then the model order is also the tradeoff between ac-
building, except for about 20% increase of occupancy between 19:00 to curacy and complexity.
21:00 each day. So in this study, the occupancy was regarded as a Resistance and capacitance (RC) method has been widely used to
model building thermal response. Fig. 7 shows a simple two-node
model for a hypothetical building external wall. The purpose of the
model is to reveal how outdoor temperature To affects indoor tem-
perature Ti, therefore, in this study, the model with indoor and outdoor
temperature is a part of the whole system model. A simple model shown
in Fig. 7(a) is used to approximate the relationship [23]. The zone node
has a thermal capacitance equal to C1, and R1 represents the thermal
resistance between outdoor and indoor. It is obvious that the model is a
first order model as shown in Eq. (4).
dTi
R1 C1 + Ti = To
dt (4)
Fig. 6. Simplified scheme of the system.
A more complex model (see Fig. 7(b)) is also used to illustrate the

5
J. Zhuang et al. Building and Environment 127 (2018) 1–12

Fig. 7. RC model.

relationship between the outdoor and indoor temperature [23]. The


thermal resistance of the external wall is divided into R1 and R2, and the
thermal capacitance of the wall C2 is added as an additional model
node. The model is described by state space equations as shown in Eq.
(5) and (6).
dTi 1 Fig. 9. The block diagram of the model.
C1 = (Tow − Ti )
dt R1 (5)

dTow 1 1 system as shown in the dashed box in Fig. 8. In other words, the HVAC
C2 = (To − Tow ) + (Ti − Tow ) system with existing control system is the modeling object in this study.
dt R2 R1 (6)
Step response method can be used to determine the system transfer
Then the relationship between To and Ti can be rewritten as a second function by injecting step signal and observing the system response
order differential equation as shown in Eq. (7). In other words, the output. If the system is approximated as a first order system, it can be
model is a second order model. expressed as a first order differential equation as Eq. (8)
d 2Ti dT dy (t ) 1 K
R1 R2 C1 C2 + (R1 C1 + R2 C2 − R2 C1) i + Ti = To + y (t ) = usup (t − τ )
dt 2 dt (7) dt T T (8)
It can be seen, the order of model has been determined at the be- And the corresponding transfer function is defines as:
ginning of modeling by using RC method. However, any model is an
Ke−τs
approximation to the modeled object, and at the beginning of modeling, Gsup (s ) =
Ts + 1 (9)
the information about the order has not yet been obtained. The order of
the model cannot be determined unless the response characteristics of where K is the gain, T is the time constant, and τ is the pure lag time.
the model can be obtained after the experiment or data analysis are These three parameters can be determined by the step response method.
implemented. At the same time, the tradeoff between accuracy and As shown in Eq. (3), the transfer function Gsup(s) can be obtained by
complexity should also be considered when developing the model. keeping another two variables as zeroes. Of course, the zero means no
Therefore, at the beginning of modeling, the order of the model should deviation from equilibrium. Thus, to minimize the impact of the load
not be determined, and should be expressed as a “RC network”, as during the experiment, it is appropriate to select a cloudy or rainy day
shown in Fig. 7(c). to do the experiment. Therefore August 1, 2016 was selected because it
was a rainy day. The difference between the highest and lowest tem-
3.4. Modeling under manipulated variables perature of the day was 4 °C. The solar radiation was small and had
little change during the daytime. The influence from these two factors
The manipulated variable is the supply chilled water temperature. If were neglected as the fluctuation of the bypass valve opening level
the variable changes, the bypass valve opening level must change. As caused by these two factors was small. Test result is presented as
shown in the dashed box of Fig. 8, if the load is constant, given a rise in Fig. 10. The step excitation was generated by setting the supply chilled
the supply chilled water temperature, the AHUs' valve opening level water temperature set point.
will become larger, and the FCs' valve will open for longer time. It The transfer function can be obtained graphically [40]. The curve of
causes the chilled water differential pressure to become smaller. In the bypass valve opening level is smoothed by least square method
order to maintain the differential pressure at the set value of 0.12 MPa, firstly as in Fig. 11. By observing, the output curve is S-shaped, so the
the controller will adjust the bypass valve to make the opening level system is a high order system. It is appropriate in engineering that the
smaller. transfer function is approximated as a first order model. The gain K can
This paper studies the relationship between the supply chilled water be calculated directly from the steady-state values of the input and
temperature and the bypass valve opening level. Taking into account output, and τ and T can be determined by plotting as shown in Fig. 11.
the short time in experiment, the load changes can be ignored, and the A tangent line is drawn for response curve, which intersects the two
block diagram of the model is shown in Fig. 9. The block in the figure steady state line. The time between the two intersections is time con-
represents the HVAC system that contains the closed loop control stant T. Pure lag time can be determined by the time between the first

Fig. 8. The control block diagram of the system.

6
J. Zhuang et al. Building and Environment 127 (2018) 1–12

Fig. 10. The step response test result.

Suu (jω) = F (Ruu (τ )) (11)

where F (•) denotes the Fourier transform, Ruu (τ ) is the auto-correlation


function of the input signal, and Ryu (τ ) is the cross-correlation function
between the output and the input signals. In the calculation of the two
correlation function, some irrelevant variables such as measurement
error, noise, etc. are suppressed, because the so-call irrelevant is that
the cross-correlation function between two variables is zero. In other
words, the Ruu (τ ) and Ryu (τ ) do not include the information of irrele-
vant variables such as measurement error and noise.
Then, Eq. (12) is employed to obtain the transfer function. The
principle is that when a random signal passes through a linear system,
the ratio between the CPSD Syu (jω) and the PSD Suu (jω) is just the
system transfer function. When the complex frequency is defined as
s = jω, the transfer function in the Laplace transform G(s) must be
equal to the one in the Fourier transform. G(jω) [45]. It is interesting to
Fig. 11. The transfer function determination by plotting. find that the formula of the transfer function for random variables is
similar to Eq. (1) for deterministic signals.
intersection and the starting point of the step response. In the three step
Syu (jω)
responses in Fig. 10, it is untrusted at 9:20, for the initial value of 0% G (jω) =
did not reflect the operating condition of the HVAC system, but in- Suu (jω) (12)
dicates that it is turned off. The model parameters estimated by two
Unlike manipulated variables, modeling with random variables re-
experiments at 14:00 and 17:30 are similar, and the average of two
quires a significant number of data samples. Data quality is as im-
identifications is taken as the model parameters. Finally, the transfer
portant as the data quantity when implementing a data-driven based
function is:
method. The data for the transition process is not desirable, for ex-
7.03e−14.9s ample, the data obtained during the HVAC system startup in the
Gsup (s ) =
84.25s + 1 morning and shutdown in the evening, since from Eq. (2), modeling
with random variables requires that the manipulated variable does not
change, and the transition process means the manipulated variable
3.5. Modeling for random variables by PSD method changed. Data of August 20–22, 2016 were selected. The sampling in-
terval was 30 s, and the data was obtained from 10:00 to 21:00 when
It is obvious that outdoor temperature and solar radiation can not be the HVAC system is under operation. During the experiment, the supply
step changed, and the step response method is not applicable for these chilled water temperature was maintained at the equilibrium value
two random variables. For random variables, a stochastic process 7 °C.
modeling method should be used. In this study, power spectral density In this study, the output is bypass valve opening level and the inputs
(PSD) is chosen for modeling, since it has the ability to suppress irre- are outdoor temperature and solar radiation. As given in Eq. (12), the
levant signals, and to improve the modeling accuracy. CPSD of output and input, and PSD of input should be estimated, re-
Firstly, the cross power spectral density (CPSD) Syu (jω) between the spectively. Then the quotient of the CPSD and the PSD will also be
output and the input signal, and the PSD Suu (jω) of the input signal are calculated. The results of the quotient for different inputs are presented
acquired, and they are the Fourier transform of the corresponding in terms of amplitude and phase at different frequencies in Figs. 12 and
correlation function as Eqs. (10) and (11). In this study, the output 13. A discrete filter is applied to fit the frequency response data by
signal y is the bypass valve opening level, and the input signal u is using least squares method [44]. And the result will be a form where
outdoor temperature or solar radiation. both the numerator and the denominator are polynomials, which is
usually the form of the transfer function. Considering the tradeoff be-
Syu (jω) = F (Ryu (τ )) (10) tween simplification and accuracy, second-order model were chosen to

7
J. Zhuang et al. Building and Environment 127 (2018) 1–12

Fig. 12. Transfer function amplitude and phase for outdoor


temperature.

obtain the transfer functions given as: inputs are determined, they are transformed to discrete form for sub-
sequent use. Secondly, these transfer functions are converted into the
2.125s + 1.268
Gout (s ) = form of state equations. Then the form of the cost function for MPC can
354.2s 2 + 42.31s + 1
be determined. The objective of this study is to minimize the opening
0.002352s + 0.01451 level of the bypass valve. Therefore, the following cost function can be
Gsol (s ) = obtained.
213.2s 2 + 25.42s + 1
N

4. Control system design


J (k ) = ∑ [w (k + i) − y (k + i k )]2
i=0 (13)

Firstly, as the three transfer functions which are related to the three subject to:

Fig. 13. Transfer function amplitude and phase for solar radiation.

8
J. Zhuang et al. Building and Environment 127 (2018) 1–12

xk + i + 1 k = Axk + i k + Buk + i k + Ddk + i k , ∀ i = 0, …, N − 1 (14) one of the sequence y*(k) is implemented on the process. As new in-
formation becomes available at the next sampling time, the above
yk + i k = Cxk + i k (15) process will be repeated. Then a new sequence is generated, and the
first value of the new sequence is used to control.
u min ≤ uk + i k ≤ u max (16) The cost function in Eq. (13) minimizes a sum of the deviations of
the bypass valve opening level from the set point, which means the
where N is the prediction horizon, the symbol uk + i k denotes the vari-
control target will be achieved. And the control structure of the system
able u at time k+i predicted at time k, u is a vector of the control inputs
is shown in Fig. 15. In the figure, random variables are outdoor tem-
within the prediction horizon N, that is the supply chilled water tem-
perature and solar radiation, and they directly act on the prediction
perature as a scalar in this study, dk + i k are the random variables, out-
model and play the role of the feedforward. The prediction model,
door temperature and solar radiation where the predicted value can be
which is composed of the three transfer functions described above,
obtained from the local weather forecast station and it represents the
predicts the future value of bypass valve opening level based on the two
feedforward control structure, yk + i k is the predicted output at time k,
which is obtained by iteratively solving Eqs. (14) and (15) using the random variables and the supply chilled water temperature. The pre-
control inputs and random variables, u min and u max are set to 7 °C and dicted supply chilled water temperature enters the controller as a
16 °C respectively. feedback value, and then the controller, according to the optimization
As discussed in section 3.1, the smaller the bypass valve opening control algorithm, adjusts the supply chilled water temperature to
level is, the higher the supply chilled water temperature is, which control the bypass valve opening level to the setpoint.
means higher chiller efficiency. However, it means that the cooling
capacity is less than the requirement of the building cooling load when 5. Results and discussions
the bypass valve is completely closed (i.e., the opening level is 0%). So
the opening level set point w (k + i) is set to be 10% to avoid the fully 5.1. Model validation
closed bypass valve.
The output variable is the controlled variable, also known as the The performance of the model under random variables is evaluated
control objective. When HVAC system energy saving is studied in this by the normalized root mean squared error (NRMSE) fitness value:
paper, chiller profile should be incorporated into system modeling if n
electricity consumption is chosen as the control variable. This would 1 ∑1 (y (k ) − yˆ (k ))2
NRMSE =
increase the model complexity and therefore electricity consumption is ymax − ymin n (17)
not adopted as the control output variable in this research. The flow
rate of supply chilled water is approximately constant due to the con- where y is the real output value and ŷ is the calculated output value. If
stant speed pumps. Therefore, energy saving can be realized by in- the matching between the calculated value and the real value is too
creasing the supply water temperature while maintaining the zone low, the parameters of the PSD method should be changed, and then the
comfort. There is no need to obtain the quantitative relationship be- PSD method is re-applied to modeling until a satisfactory matching
tween the supply chilled water and chiller energy consumption because degree is obtained. The parameters include samples of overlap from
such relationship is not needed in the optimization calculation. section to section and the selected window function. After sensitivity
In Eq. (13), the supply water temperature is not used as a controlled test, parameters for PSD were set as follows: the window is 10 that
variable. However, in this study, the bypass valve opening level is se- means a Hamming window of equal length is used, and the overlap is
lected as the controlled variable as shown in Fig. 14. Here the HVAC the default value 50%.
system with the existing control system is demonstrated in the dash box The bypass valve opening level calculated by the weather data of
in Fig. 8 and is illustrated in Fig. 9. Although the bypass valve opening August 23, 2016 is compared to the real opening level, as shown in
level is a manipulate variable in the original system as shown in Fig. 8, Fig. 16. NRMSE is 0.127 meaning that the model has a fitness of 87.3%.
it is used as a controlled variable as shown in Fig. 14 for the following Fig. 17 demonstrates the discrepancy between the calculated value Yp
reason. If the supply water temperature is selected as the controlled and the real value Yr, all the blue point are clustered around the line:
variable, the control structure in Fig. 14 has to be changed in Eq. (14) as Yr = Yp. From these two figures, it can be seen that the model accuracy
a constraint condition. Because in order to ensure the indoor comfort, is satisfied. The comparison does not include the transition process
the supply water temperature cannot increase without limit, in other because the transition process after chillers re-start is not modeled in
words, the bypass valve opening level must not be closed. Solving this this study.
constraint is same as the handling the opening level when it is used as
the controlled variable, therefore, can also be solved through Eqs. 5.2. Control simulation
(13)–(16).
The prediction horizon, N is normally selected as large as possible in Before incorporating the MPC strategy in an online experiment, it
the premise of ensuring the computation ability. In this study, the was tested under simulation environment. The objective of the simu-
HVAC system is turned off in every night. Thus, the prediction horizon lation study was to examine the closed loop performance. Usually,
at each sample time is variable. In each sampling time, the optimal when MPC is applied online, the control object outputs are measured by
sequence {y∗ (k + i)}iN=−01 is obtained by using Eqs. (13)–(16). The first the sensors and then fed back to the controller. For simulation purpose,

Fig. 14. The control structure of this study.

9
J. Zhuang et al. Building and Environment 127 (2018) 1–12

Fig. 15. The block diagram of the MPC with feedforward


structure.

CC CC ⎞2
Ih = ⎧I0⋅sinh⋅⎡C0 + C1⋅ + C2⋅⎛ + C3⋅(θn − θn − 3) + C4 ϕ ⎤ − C5 ⎫
⎨ ⎢
⎣ 10 ⎝ 10 ⎠ ⎥
⎦ ⎬
⎩ ⎭
/k (18)

Where Ih is total horizontal radiation, I0 is solar constant and the value


is 1353 W/m2, h is solar elevation angel, CC is cloud cover and the
value is 0,1, …10, θn and θn − 3 are the temperatures at some moment
and 3 h ago, ϕ is relative humidity, and C0…C5 and k are constants
given in Table 1. Fig. 18 demonstrates the simulation result from test
data obtained in August 24, 2016. Since the transition process after
chillers re-start is not studied, the control is implemented when system
reaches a steady state.
Fig. 16. Comparison of the calculated and real value.

5.3. Control experiment

After simulation, the data in same day is also used in real test.
Fig. 18 (a) and (b) show the bypass valve opening level and the supply
chilled water temperature in the simulation and the actual experiment
respectively. The results are very similar. It shows that MPC can over-
come the model mismatch caused by simplification in modeling pro-
cess. MPC algorithm includes three parts: prediction model, rolling
optimization and feedback correction, and the latter two ensure that the
control system has a feedback structure, which can tolerate a certain
degree of model mismatch.
The electricity consumption in August 23, 2016 and August 24,
2016 were used to evaluate the energy saving potential by increasing
supply chilled water temperature. The traditional control method was
implemented in August 23, which from 9:00 to 21:00, the average
outdoor temperature and solar radiation, and the electricity consump-
tion of the two days are listed in Table 2. It can be seen that two days of
climate is almost the same. And it can obtain a 16% chiller energy
saving that the control method for increasing supply chilled water
temperature is applied, compared with the traditional method that
Fig. 17. Bypass valve opening level prediction error. maintain the temperature at 7 °C. In fact, if the outdoor temperature is
lower than the two experiment days, the energy saving potential is
the controlled object in Fig. 15 is replaced by the model same as the greater due to the higher supply water temperature that can meet the
prediction model. Although the prediction model is different from the lower load. Increasing the supply chilled water temperature may affect
real system, MPC can still overcome the small model mismatch, since the dehumidification effect. And the indoor relative humidity is also
MPC algorithm includes a feedback alignment. listed in Table 2. It can be seen that the humidity is slightly higher by
The weather prediction only have the maximum and minimum
Table 1
outdoor temperatures for the next day, and the solar radiation has no
Constant value used in Eq. (18) [42].
prediction. Outdoor temperature forecast can be replaced by sinusoidal
curve [41]. Solar radiation prediction can be obtained by Eq. (18) [42]: C0 C1 C2 C3 C4 C5 k

0.3645 0.48 −0.6335 0.0495 −0.0011 40.566 0.9900

10
J. Zhuang et al. Building and Environment 127 (2018) 1–12

Fig. 18. Comparison results of MPC strategy between simulation


and experiment.

Table 2 and PSD method are used respectively for system modeling under these
Comparison results of the electricity consumption in two days. two categories of variables. Step response method is more applicable in
the real practice due to its simplicity and easy to implement. PSD
Date Avg. outdoor Avg. solar Electricity Indoor relative
temperature radiation consumption humidity method belongs to data driven methods, and its accuracy depends on
the richness of the data. Given the proposed method, higher model
Aug. 23 35.4 °C 1145 W/m2 7337 kWh 49.3%–53.5% accuracy is achieved and modeling complex is lowered through being
Aug. 24 35.1 °C 1089 W/m2 6161 kWh 62.7%–68.7%
applied in a medium sized commercial building in China.
The test bypass valve in chiller plant was controlled through the
proposed MPC strategy. Energy saving can be achieved by the proposed
using the control method of this study, due to increased supply chilled
MPC strategy.
water temperature and dehumidification capacity decreased. Although,
We expect to further improve the accuracy of this bypass valve
humidity control is critical to maintain thermal comfort, the range of
control model by including more random variables such as occupancy
relative humidity in this building is set to 55%–70% which can be
and electricity rate. It is worth studying that how much of these vari-
maintained even the chilled water supply temperature fluctuates in
ables that will be added to improve the accuracy of the model, and then
certain scope. Therefore, in this research, the humidity control was not
improve the control performance.
considered.

Acknowledgements
6. Conclusion and future work
The research work presented in this paper is financially supported
Buildings are usually approximated as linear systems, and there is a by Beijing Municipal Education Commission science and technology
tradeoff between the model accuracy and the control performance. If a program [Grant No. SQKM201510016002].
control strategy can well compensate model error, the model can be
simplified to a low order model, vice versa. In the real practice, the References
error caused by the approximation of a high-order system to a low-
order system is allowable, and the approximation can greatly simplify [1] C.B. Zhou, S.J. Sheng, Study on integrated design of passive energy saving in hot
summer and warm winter region residential building, J. FUZHOU Univ. Nat. Sci.
the analysis of modeling. This study proposes a new modeling method Ed. 39 (5) (2011) 760–764 (in Chinese).
for a medium-sized commercial building. A MPC strategy with feed- [2] Z.F. Xue, Y. Jiang, Study on the air conditioning energy consumption guideline in
forward control structure is developed to increase the supply chilled commercial buildings, HV AC 35 (1) (2005) 37–41 (in Chinese).
[3] A. Afram, F. Janabi-Sharifi, Theory and applications of HVAC control systems – a
water temperature. review of model predictive control (MPC), Build. Environ. 72 (1) (2014) 343–355.
In the proposed modeling framework, the inputs are divided into [4] E. Khanmirza, A. Esmaeilzadeh, A.H.D. Markazi, Predictive control of a building
two categories as manipulated variable and random variable to analyze hybrid heating system for energy cost reduction, Appl. Soft Comput. 46 (2016)
407–423.
their effects on the system by different methods. Step response method

11
J. Zhuang et al. Building and Environment 127 (2018) 1–12

[5] S. Salakij, N. Yu, S. Paolucci, P. Antsaklis, Model-Based Predictive Control for [24] Y. Ma, A. Kelman, A. Daly, F. Borrelli, Predictive control for energy efficient
building energy management. I: energy modeling and optimal control, Energy buildings with thermal storage: modeling, stimulation, and experiments, IEEE
Build. 133 (2016) 345–358. Control Syst. 32 (1) (2012) 44–64.
[6] N. Yu, S. Salakij, R. Chavez, S. Paolucci, M. Sen, P. Antsaklis, Model-based pre- [25] J.E. Braun, N. Chaturvedi, An inverse gray-box model for transient building load
dictive control for building energy management: Part II–Experimental validations, prediction, HVAC & R Res. 8 (1) (2002) 73–99.
Energy Build. 146 (2017) 19–26. [26] J. Hu, P. Karava, A state-space modeling approach and multi-level optimization
[7] H. Huang, L. Chen, E. Hu, A new model predictive control scheme for energy and algorithm for predictive control of multi-zone buildings with mixed-mode cooling,
cost savings in commercial buildings: an airport terminal building case study, Build. Build. Environ. 80 (7) (2014) 259–273.
Environ. 89 (2015) 203–216. [27] R.C. Dorf, R.H. Bishop, Modern Control Systems, eleventh ed., Pearson Prentice
[8] A.E. Ruano, S. Pesteh, S. Silva, H. Duarte, G. Mestre, P.M. Ferreira, H.R. Khosravani, Hall, 2008.
R. Horta, The IMBPC HVAC system: a complete MBPC solution for existing HVAC [28] D.E. Seborg, T.F. Edgar, D.A. Mellichamp, F.J. Doyle III, Process Dynamics and
systems, Energy Build. 120 (2016) 145–158. Control, third ed., John Wiley & Sons, Inc, 2011.
[9] C.R. Cutler, B.L. Ramaker, Dynamic matrix control – a computer control algorithm, [29] E. O'Dwyer, L. De Tommasi, K. Kouramas, M. Cychowski, G. Lightbody, Modelling
Proceedings of the 1980 Joint Automatic Control Conference, Aug. 13-15, 1980 San and disturbance estimation for model predictive control in building heating sys-
Francisco. tems, Energy Build. 130 (2016) 532–545.
[10] S.J. Qin, T.A. Badgwell, A survey of industrial model predictive control technology, [30] G. Fraisse, C. Viardot, O. Lafabrie, G. Achard, Development of a simplified and
Control Eng. Pract. 11 (7) (2003) 733–764. accurate building model based on electrical analogy, Energy Build. 34 (10) (2002)
[11] X. Li, J. Wen, Building energy consumption on-line forecasting using physics based 1017–1031.
system identification, Energy Build. 82 (2014) 1–12. [31] Y. Yao, K. Yang, M. Huang, L. Wang, A state-space model for dynamic response of
[12] B.C. Kuo, F. Golnaraghi, Automatic Control Systems, eighth ed., John Wiley & Sons, indoor air temperature and humidity, Build. Environ. 64 (2013) 26–37.
Inc., 2003. [32] P.R. Armstrong, S.B. Leeb, L.K. Norford, Control with building mass – Part I:
[13] A. Afram, F. Janabi-Sharifi, Review of modeling methods for HVAC systems, Appl. thermal response model, ASHRAE Trans. (2006) 112.
Therm. Eng. 67 (1–2) (2014) 507–519. [33] N.T. Gayeski, P.R. Armstrong, L.K. Norford, Predictive pre-cooling of thermo-active
[14] A. Afram, F. Janabi-Sharifi, A.S. Fung, K. Raahemifar, Artificial neural network building systems with low-lift chillers, HVAC & R Res. 18 (5) (2012) 858–873.
(ANN) based model predictive control (MPC) and optimization of HVAC systems: a [34] B. Huchuk, C.A. Cruickshank, W. O'Brien, H.B. Gunay, Recursive Thermal Building
state of the art review and case study of a residential HVAC system, Energy Build. Model Training Using Ensemble Kalman Filters, Ottowa, Canada, eSim, (2014).
141 (2017) 96–113. [35] L. Ferkl, J. Široky, Ceiling radiant cooling: comparison of ARMAX and subspace
[15] H. Huang, L. Chen, E. Hu, A neural network-based multi-zone modelling approach identification modelling methods, Build. Environ. 45 (1) (2010) 205–212.
for predictive control system design in commercial buildings, Energy Build. 97 [36] P. Defabianis, A. Cimma, D. Tessaris, R. Lala, L. Basano, P. Appendino,
(2015) 86–97. E. Caravaggio, Integration of fault detection and diagnosis with energy manage-
[16] P. Ferreira, A. Ruano, S. Silva, E. Conceicao, Neural networks based predictive ment standard ISO 50001 and operations and maintenance of HVAC systems, Rehva
control for thermal comfort and energy savings in public buildings, Energy Build. World Congress and International Conference on Iaqvec Clima, 2013.
55 (2012) 238–251. [37] D.S. Carrasco, G.C. Goodwin, Feedforward model predictive control, Annu. Rev.
[17] J. Sun, A. Reddy, Optimal control of building HVAC & R systems using complete Control 35 (2) (2011) 199–206.
simulation-based sequential quadratic programming (CSB-SQP), Build. Environ. 40 [38] J.A. Rossiter, G. Valencia-Palomo, Feed forward design in MPC, Control Conference,
(5) (2005) 657–669. 2009.
[18] B. Tashtoush, M. Molhim, M. Al-Rousan, Dynamic model of an HVAC system for [39] E.J. Davison, The feedforward control of linear multivariable time-invariant sys-
control analysis, Energy 30 (10) (2005) 1729–1745. tems, Automatica 9 (5) (1973) 561–573.
[19] F. Tahersima, J. Stoustrup, H. Rasmussen, P.G. Nielsen, Thermal analysis of an [40] B.W. Bequette, Process Dynamics: Modeling, Analysis and Simulation, Prentice
HVAC system with TRV controlled hydronic radiator, Automation Sci. Eng. 19 (11) Hall, 1998.
(2010) 756–761. [41] P.C. Tabares-Velasco, C. Christensen, M. Bianchi, Verification and validation of
[20] M. Sourbron, C. Verhelst, L. Helsen, Building models for model predictive control of EnergyPlus phase change material model for opaque wall assemblies, Build.
office buildings with concrete core activation, J. Build. Perform. Simul. 6 (3) (2012) Environ. 54 (7) (2012) 186–196.
1–24. [42] Q.Y. Zhang, H.X. Yang, Typical Meteorological Database Handbook for Buildings,
[21] R. Bălan, J. Cooper, K.-M. Chao, S. Stan, R. Donca, Parameter identification and China Architecture & Building Press, 2012 (in Chinese).
model based predictive control of temperature inside a house, Energy Build. 43 [43] Y. Ma, Model Predictive Control for Energy Efficient Buildings, PhD Dissertations
(2–3) (2011) 748–758. University of California, Berkeley, 2012.
[22] H. Karlsson, C.-E. Hagentoft, Application of model based predictive control for [44] L. He, Z. Jin, The population predicting based on the curve fitting least square
water-based floor heating in low energy residential buildings, Build. Environ. 46 (3) method, International Conference on Automation, Mechanical Control and
(2011) 556–569. Computational Engineering, 2015.
[23] S. Prívara, J. Široký, L. Ferkl, J. Cigler, Model predictive control of a building [45] A.V. Oppenheim, A.S. Willsky, I.T. Young, Signals and Systems, Prentice-Hall Single
heating system: the first experience, Energy Build. 43 (2–3) (2011) 564–572. Processing Series Prentice-Hall, Englewood Cliffs, 1983.

12