File preview
Methods
for
Network
Enabled
Embedded
Monitoring
and
Control
for
High-‐Performance
Buildings
Prabir
Barooah
(UF),
Prashant
G.
Mehta
&
Sean
Meyn
(UIUC),
Luca
Carloni
(Columbia
U),
Martin
Krucinski
(UTRC) Acknowledgement:
NSF
grant
CNS-‐0931885 (March
2010)
✦
Occupancy
modeling
and
prediction,
control
of
evacuation Platform-‐based
design
space
exploration
for
cyber
systems
relevant
to
building
monitoring
and
control Energy
efficiency
improvement
through
estimation
and
control
Learning
and
adaptation
of
models
and
networks
across
buildings
and
time-‐scales
✦
✦
✦
✦
Occupancy
modeling
and
prediction,
control
of
evacuation Platform-‐based
design
space
exploration
for
cyber
systems
relevant
to
building
monitoring
and
control Energy
efficiency
improvement
through
estimation
and
control
Learning
and
adaptation
of
models
and
networks
across
buildings
and
time-‐scales
✦
✦
✦
Buildings and Energy (U.S.A.)
>
70%
of
electricity*,
40%
of
total
energy,
... >
commercial
vs.
residential
:
close
to
1
Sources
of
Inefficiency:
(1)
over-‐design,
(2)
poor
operation
HVAC
~
50%
heating,
ventilation,
air
conditioning 2050
U.S.
building
energy
use
BAU with savings
State of the art?
inefficiency:
(1)
over-‐design
(2)
poor
operation
90 80 Temp
(F) 70 60 50 0 hour-‐-‐-‐> 5 10 15 noon 20 serve
the
occupants,
not
empty
space! Air
quality!!!
(ASHRAE
standards) Supply
air Room Pugh
Hall,
Univ
of
Florida
Room
249,
Pugh
Hall
High performance buildings
efficiency:
(1)
...
(2)
better
operation
“minimize
energy
use
while
maintaining
comfort
and
IAQ”
weather
model occupancy
model
optimizer temp
+
air
quality model min
E
subj.
to
....
complexity
and
uncertainty
makes
modeling/prediction/design
challenging *)
Delay
between
action
and
its
effect *)
Air
quality
and
temp
vs.
energy Occupancy
prediction
none
some
perfect
Predictive control w/o occp. prediction
Set-‐point
based
control
(PID+logic) Floating-‐point
control: occupied T-‐des
T-‐des
*)
allow
temp
within
a
band un-‐occupied *)
wider
band
when
unoccupied *)
maintain
minimum
airflow
to
remove
VOC
and
CO2
Formulated
as
a
RHC
*)
if
occupied
now,
assume
will
be
(Receding
horizon
optimal
control)
occupied
for
the
next
T
minutes problem *)
control
changes
only
every
T
minutes *)
minimize
energy
over
the
next
NT
minutes
Example
76
Room
Temperature
0.4
Power
consumption PID
RHC
74
0F
0.3
72
RHC PID
70
68
hour
-‐-‐-‐-‐>
6 12 20 24
kW
0.2 0.1 0
4
4
6
12
20
24
occupancy
occupancy
30%
savings
over
PID (ex
2:
occupied
18
of
24
hours:
10%)
Example
76
Room
Temperature
0.4
Power
consumption PID
RHC
74
0F
0.3
72
RHC PID
70
68
hour
-‐-‐-‐-‐>
6 12 20 24
kW
0.2 0.1 0
4
4
6
12
20
24
occupancy
occupancy
30%
savings
over
PID (ex
2:
occupied
18
of
24
hours:
10%)
✦
Large
savings
are
possible
with
only
occupancy
measurement,
w/o
prediction bottleneck
is
not
temperature
control
but
air
quality/ventilation
(health
problems,
lawsuits,
..)
✦
Example
76
Room
Temperature
0.4
Power
consumption PID
RHC
74
0F
0.3
72
RHC PID
70
68
hour
-‐-‐-‐-‐>
6 12 20 24
kW
0.2 0.1 0
4
4
6
12
20
24
occupancy
occupancy
30%
savings
over
PID (ex
2:
occupied
18
of
24
hours:
10%)
✦
Large
savings
are
possible
with
only
occupancy
measurement,
w/o
prediction bottleneck
is
not
temperature
control
but
air
quality/ventilation
(health
problems,
lawsuits,
..)
✦
Measuring occupancy
Most
sensors
are
“flow”
sensors
=>
large
drift! Occp(k)
=
flow
in(1)
+
flow
in(2)
+
.....
+
flow
in(k)
200
data
from
Pugh
Hall
@UF
150
#
of
occupants
100
50
0
−50 0
6
hour
of
day
10 12
15
20
24
Measuring occupancy
Most
sensors
are
“flow”
sensors
=>
large
drift! Occp(k)
=
flow
in(1)
+
flow
in(2)
+
.....
+
flow
in(k)
200
12
data
from
Pugh
Hall
@UF
10
experiment
in
a
6
occupant
office Software
sensor
(video) TRUE CPS
150
#
of
occupants
#
of
occupants
6 10 12 15 20 24
8 6 4 2 0 0
100
50
0
CO2
−50 0
hour
of
day
8
12
16
20
24
Measuring occupancy
Most
sensors
are
“flow”
sensors
=>
large
drift! Occp(k)
=
flow
in(1)
+
flow
in(2)
+
.....
+
flow
in(k)
200
12
data
from
Pugh
Hall
@UF
10
experiment
in
a
6
occupant
office Software
sensor
(video) TRUE CPS
150
#
of
occupants
#
of
occupants
6 10 12 15 20 24
8 6 4 2 0 0
100
50
0
CO2
−50 0
hour
of
day
8
12
16
20
24
Cyber
Physical
Sensor: software
to
detect
mouse/keyboard
idle
Exploit inter-zone convection
Which
rooms
have
convection
interaction
(airflow
thermal
exchange)? A
network
structure
identification
problem
proposed
solution
based
on
estimating
conditional
probabilities
Room 243 78
76
Temperature(F)
74
72
70
68 0
10
20
Time (hr)
30
Without Convection With Convection Actual 40 50
Test-bed Development
✦
BACnet
stack
server
✦
Pugh
Hall
@UF
campus LEED
Silver,
66
zones
runs
on
a
PC
residing
@Pugh
communication
through
VLAN
UF Campus Map
✦
✦
Capable
of
high
frequency
control
of
all
actuators
(not
just
“set
point”
change) Control
software
being
stress-‐ tested
http://ca
Campus Map
University of Flo
Search
BUILDING LOCATOR MAP FEATURES
✦
Art, Culture & History Campus Life Sports & Recreation Sustainable Campus Transportation Technology CALCULATE DISTANCE
Welcome to the UF Campus Map!
PPD
Since these maps are based on the Google maps and are draggable, you can use your mouse or the directional arrows to pan left, right, up and down to see areas that are hidden offscreen. You can also use the slider to zoom in and zoom out. Single-clicking on a map feature (such as a building
Team
UF Prabir
Barooah Sean
Meyn UIUC Prashant
Mehta Columbia United
Tech.
Res.
Center Luca
Carloni Martin
Krucinski
Siddharth
Goyal Chenda
Liao Yashen
Lin
Kun
Deng
Adam
Tilton
Marcin
Szczodrak
Collaborators: Sanjay
Ranka,
faculty,
CISE@UF (cyber
physical
sensor)
Tim
Middelkoop,
faculty.,
ISE@UF (Pugh
hall
DAC
server)
for
Network
Enabled
Embedded
Monitoring
and
Control
for
High-‐Performance
Buildings
Prabir
Barooah
(UF),
Prashant
G.
Mehta
&
Sean
Meyn
(UIUC),
Luca
Carloni
(Columbia
U),
Martin
Krucinski
(UTRC) Acknowledgement:
NSF
grant
CNS-‐0931885 (March
2010)
✦
Occupancy
modeling
and
prediction,
control
of
evacuation Platform-‐based
design
space
exploration
for
cyber
systems
relevant
to
building
monitoring
and
control Energy
efficiency
improvement
through
estimation
and
control
Learning
and
adaptation
of
models
and
networks
across
buildings
and
time-‐scales
✦
✦
✦
✦
Occupancy
modeling
and
prediction,
control
of
evacuation Platform-‐based
design
space
exploration
for
cyber
systems
relevant
to
building
monitoring
and
control Energy
efficiency
improvement
through
estimation
and
control
Learning
and
adaptation
of
models
and
networks
across
buildings
and
time-‐scales
✦
✦
✦
Buildings and Energy (U.S.A.)
>
70%
of
electricity*,
40%
of
total
energy,
... >
commercial
vs.
residential
:
close
to
1
Sources
of
Inefficiency:
(1)
over-‐design,
(2)
poor
operation
HVAC
~
50%
heating,
ventilation,
air
conditioning 2050
U.S.
building
energy
use
BAU with savings
State of the art?
inefficiency:
(1)
over-‐design
(2)
poor
operation
90 80 Temp
(F) 70 60 50 0 hour-‐-‐-‐> 5 10 15 noon 20 serve
the
occupants,
not
empty
space! Air
quality!!!
(ASHRAE
standards) Supply
air Room Pugh
Hall,
Univ
of
Florida
Room
249,
Pugh
Hall
High performance buildings
efficiency:
(1)
...
(2)
better
operation
“minimize
energy
use
while
maintaining
comfort
and
IAQ”
weather
model occupancy
model
optimizer temp
+
air
quality model min
E
subj.
to
....
complexity
and
uncertainty
makes
modeling/prediction/design
challenging *)
Delay
between
action
and
its
effect *)
Air
quality
and
temp
vs.
energy Occupancy
prediction
none
some
perfect
Predictive control w/o occp. prediction
Set-‐point
based
control
(PID+logic) Floating-‐point
control: occupied T-‐des
T-‐des
*)
allow
temp
within
a
band un-‐occupied *)
wider
band
when
unoccupied *)
maintain
minimum
airflow
to
remove
VOC
and
CO2
Formulated
as
a
RHC
*)
if
occupied
now,
assume
will
be
(Receding
horizon
optimal
control)
occupied
for
the
next
T
minutes problem *)
control
changes
only
every
T
minutes *)
minimize
energy
over
the
next
NT
minutes
Example
76
Room
Temperature
0.4
Power
consumption PID
RHC
74
0F
0.3
72
RHC PID
70
68
hour
-‐-‐-‐-‐>
6 12 20 24
kW
0.2 0.1 0
4
4
6
12
20
24
occupancy
occupancy
30%
savings
over
PID (ex
2:
occupied
18
of
24
hours:
10%)
Example
76
Room
Temperature
0.4
Power
consumption PID
RHC
74
0F
0.3
72
RHC PID
70
68
hour
-‐-‐-‐-‐>
6 12 20 24
kW
0.2 0.1 0
4
4
6
12
20
24
occupancy
occupancy
30%
savings
over
PID (ex
2:
occupied
18
of
24
hours:
10%)
✦
Large
savings
are
possible
with
only
occupancy
measurement,
w/o
prediction bottleneck
is
not
temperature
control
but
air
quality/ventilation
(health
problems,
lawsuits,
..)
✦
Example
76
Room
Temperature
0.4
Power
consumption PID
RHC
74
0F
0.3
72
RHC PID
70
68
hour
-‐-‐-‐-‐>
6 12 20 24
kW
0.2 0.1 0
4
4
6
12
20
24
occupancy
occupancy
30%
savings
over
PID (ex
2:
occupied
18
of
24
hours:
10%)
✦
Large
savings
are
possible
with
only
occupancy
measurement,
w/o
prediction bottleneck
is
not
temperature
control
but
air
quality/ventilation
(health
problems,
lawsuits,
..)
✦
Measuring occupancy
Most
sensors
are
“flow”
sensors
=>
large
drift! Occp(k)
=
flow
in(1)
+
flow
in(2)
+
.....
+
flow
in(k)
200
data
from
Pugh
Hall
@UF
150
#
of
occupants
100
50
0
−50 0
6
hour
of
day
10 12
15
20
24
Measuring occupancy
Most
sensors
are
“flow”
sensors
=>
large
drift! Occp(k)
=
flow
in(1)
+
flow
in(2)
+
.....
+
flow
in(k)
200
12
data
from
Pugh
Hall
@UF
10
experiment
in
a
6
occupant
office Software
sensor
(video) TRUE CPS
150
#
of
occupants
#
of
occupants
6 10 12 15 20 24
8 6 4 2 0 0
100
50
0
CO2
−50 0
hour
of
day
8
12
16
20
24
Measuring occupancy
Most
sensors
are
“flow”
sensors
=>
large
drift! Occp(k)
=
flow
in(1)
+
flow
in(2)
+
.....
+
flow
in(k)
200
12
data
from
Pugh
Hall
@UF
10
experiment
in
a
6
occupant
office Software
sensor
(video) TRUE CPS
150
#
of
occupants
#
of
occupants
6 10 12 15 20 24
8 6 4 2 0 0
100
50
0
CO2
−50 0
hour
of
day
8
12
16
20
24
Cyber
Physical
Sensor: software
to
detect
mouse/keyboard
idle
Exploit inter-zone convection
Which
rooms
have
convection
interaction
(airflow
thermal
exchange)? A
network
structure
identification
problem
proposed
solution
based
on
estimating
conditional
probabilities
Room 243 78
76
Temperature(F)
74
72
70
68 0
10
20
Time (hr)
30
Without Convection With Convection Actual 40 50
Test-bed Development
✦
BACnet
stack
server
✦
Pugh
Hall
@UF
campus LEED
Silver,
66
zones
runs
on
a
PC
residing
@Pugh
communication
through
VLAN
UF Campus Map
✦
✦
Capable
of
high
frequency
control
of
all
actuators
(not
just
“set
point”
change) Control
software
being
stress-‐ tested
http://ca
Campus Map
University of Flo
Search
BUILDING LOCATOR MAP FEATURES
✦
Art, Culture & History Campus Life Sports & Recreation Sustainable Campus Transportation Technology CALCULATE DISTANCE
Welcome to the UF Campus Map!
PPD
Since these maps are based on the Google maps and are draggable, you can use your mouse or the directional arrows to pan left, right, up and down to see areas that are hidden offscreen. You can also use the slider to zoom in and zoom out. Single-clicking on a map feature (such as a building
Team
UF Prabir
Barooah Sean
Meyn UIUC Prashant
Mehta Columbia United
Tech.
Res.
Center Luca
Carloni Martin
Krucinski
Siddharth
Goyal Chenda
Liao Yashen
Lin
Kun
Deng
Adam
Tilton
Marcin
Szczodrak
Collaborators: Sanjay
Ranka,
faculty,
CISE@UF (cyber
physical
sensor)
Tim
Middelkoop,
faculty.,
ISE@UF (Pugh
hall
DAC
server)