Thermal Comfort Control Based on Neural Network for HVAC Application

1
Thermal Comfort Control Based on Neural Network
for HVAC Application
Jian Liang and Ruxu Du Dept. of Automation and
Computer-Aided Engineering The Chinese University
of Hong Kong August 2005
2
Outline

• Introduction
• Design of the thermal comfort Controller
• Models of the thermal comfort Controller
• Design of the Neural Networks controller
• Simulation of the thermal comfort Controller
• Conclusion and further research

3
Introduction

• The Heating, Ventilating and Air Conditioning
  (HVAC) plays an important role in energy
  consumption
• In China, it takes 15 of the building energy
• In United States, it takes 44
• Development of air-conditioning control
• First generation ON / OFF switch based on the
  sensation of the users
• Second generation ON / OFF control assisted by a
  temperature sensor
• Third generation, digital control assisted by
  electronic thermostat, and humidity was also
  taken into consideration
• Fourth generation intelligent control (fuzzy
  control, adaptive control and etc.)

4
Introduction

• Background
• Most of the HVAC systems still adopt the
  temperature / humidity controllers
• Thermal comfort control is necessary for higher
  comfort level
• Thermal comfort indices
• Standard Effective Temperature (SET) (Gagge,
  1971)
• Predicted Mean Vote (PMV) (Fanger, 1970) predict
  the mean thermal sensation vote on a standard
  scale for a large group of persons
• PMV have been adopted by ISO 7730 standard, and
  ISO recommends to maintain PMV at 0 with a
  tolerance of 0.5 as the best thermal comfort
• Thermal comfort concept for long exposure to a
  constant thermal environment with a constant
  metabolic rate, a heat balance can be established
  for the human body and the bodily heat production
  is equal to its heat dissipation

5
Introduction

• Background
• Thermal comfort variables for PMV calculation
• Four environmental-dependent variables air
  temperature Ta, relative air humidity RH,
  relative air velocity Vair, mean radiant
  temperature Tmrt
• Two personal-dependent variables activity level
  , clo-value (related to clothing worn by the
  occupants)
• As a measure for the thermal comfort, one can use
  the seven point psycho-physical ASHRAE scale

6
Introduction

• Air conditioning controller
• Most of the AC controllers are air temperature
  regulator (ATR)
• These regulators control the indoor
  temperature / humidity. Since comfort level is
  determined by six variables, thus these
  regulators cant provide high comfort level
• Comfort index regulators were proposed (CIR)
  (MacArthur, 1986 Scheatzle, 1991)
• These regulators are based on PMV / SET. The
  default reference input is 0 (neutral). Occupant
  serves as a supervisory controller by adjusting
  the reference value
• User-adaptable comfort controller (UACC)
  (Federspiel and Asada, 1994 )
• These controllers are based on a simplified
  PMV-like index proposed by Federspiel. It can
  tune the PMV model parameters by learning the
  specific occupants thermal sensation.
• Some thermal comfort sensing systems were
  designed (J. Kang and S. Park, 2000)

7
Introduction

• Our objective design an intelligent thermal
  comfort controller based on neural networks for
  HVAC application
• High comfort level
• Learn the comfort zone from the users
  preference, and guarantee the high comfort level
  and good dynamic performance
• Energy saving
• Combine the thermal comfort control with a
  energy saving strategy
• Air quality control
• Provide variable air volume (VAV) control,
  and adjust the fresh air and return air mix ratio
  to guarantee the required fresh air

8
Thermal comfort controller design

• Block diagram of the thermal comfort control
  system

9
Thermal comfort controller design

• Comfort zone learning logic

10
Models of the thermal comfort controller

• Thermal sensation model
• The PMV formula proposed by Fanger (1970)
• where M metabolism (w/m2)
• W external work,
  equal to zero for most activity (w/m2)
• M metabolism
  (w/m2)
• Icl thermal
  resistance of clothing (clo)
• fcl ratio of
  bodys surface area when fully clothed to bodys
  surface area when nude
• Pa partial water
  vapor pressure (Pa)

11
Models of the thermal comfort controller

• Thermal sensation model
• The personal-dependant variables, activity level
  and the clo-value cant be measured directly, and
  hence, in the practical design, they are set as
  constant parameters according to different season
• The PMV calculation formula is nonlinear and
  necessitate iterative calculation. In the
  simulation, a computer calculation model proposed
  by D. Int-Hout is used
• If high real time performance is required, we can
  also adopt the PMV-like index (Federspiel and
  Asada, 1994)
• Or we can also use Neural Network to build a PMV
  calculation model

12
Models of the thermal comfort controller

• Thermal space model
• A lumped parameter single-zone house model is
  built
• The sensible and latent energy exchange is taken
  into consideration
• The indoor air velocity is assumed proportional
  to the input airflow rate
• A uniform wall temperature is assumed and
  regarded equal to the mean radiant temperature,
  etc.

13
Models of the thermal comfort controller

• Thermal space model
• Three input variables cooling capacity, air flow
  rate, fresh air and return air mix ratio
• Three disturbances indoor heat load, ambient
  temperature and humidity

14
Design of NN controller

• Controller design
• The conventional comfort controllers are based on
  the on-off control or PI / PID control
• To overcome the nonlinear feature of PMV
  calculation, time delay and system uncertainty,
  some advanced control algorithms have been
  proposed
• Fuzzy adaptive control (Dounis and Manolakis,
  2001 Calvino et al, 2004)
• Optimal comfort control (MacArthur and Grald,
  1988)
• Minimum-power comfort control (Federspiel and
  Asada, 1994)
• A kind of direct NN controller is designed based
  on back-propagation algorithm in this paper,
  which has been successfully applied in the
  hydronic heating systems (A. Kanarachos et al,
  1998)

15
Design of NN controller

• NN Controller design
• A two-layer MISO NN controller is designed, which
  has two inputs and one output e is the error
  between the PMV set value and feedback value, is
  the error derivative and u is the output to
  control the HVAC system.

16
Simulation of the thermal comfort controller

• I. Settings of major simulation parameters
• Heating and cooling performance are investigated
• CAV (constant-air-volume) and VAV
  (variable-air-volume) applications are
  investigated

17
Simulation of the thermal comfort controller

• II. System performance under thermal comfort
  control and
• temperature control
• For the temperature control, the reference input
  is 23oC (cooling) and 25oC (heating)
• For the comfort control, the reference input is 0

18
Simulation of the thermal comfort controller

• III. System performance under direct NN control
  and PI
• control
• For the well-tuned PI controller with integral
  anti-windup,
• When the control output reaches the
  limitation, the integral action is cut off
• For the comfort controller, the learning
  coefficient is set as ? 0.315

19
Simulation of the thermal comfort controller

• IV. Cooling / heating response under thermal
  comfort control

20
Simulation of the thermal comfort controller

• V. Minimum-power control strategy under VAV
  Control
• By adjusting the air flow rate fmix, mixed air
  ratio r, and the PMV value according to the
  users comfort zone, energy saving can be
  obtained

21
Simulation of the thermal comfort controller

• VI. System Performance under CAV and VAV Control
• Within 12 hours, cooling power consumed by VAV
  and CAV systems are 25.93KWh and 28.93KWh
  respectively, and hence, 3KWh cooling power can
  be saved

22
Conclusion and further work

• Conclusion
• The conventional temperature controller (on /
  off control or PI control ), cant guarantee the
  high comfort level (PMV 0)
• The thermal comfort controller can keep the
  thermal environment at the highest level
• The designed NN controller has good control
  performance and disturbance rejection ability,
  and easy to fine tune in practice
• The proposed minimum-power control strategy can
  achieve high comfort level as well as the energy
  saving at the same time
• Further work
• Measurement of the activity level and the
  clo-value
• Location of sensor
• Development of the cost-effective thermal comfort
  control system

23
Question and Answer
Thank you