Player is loading...
Predictive controls for HVAC systems
Urs Grossenbacher, Pronoó AG, Givisiez, Switzerland
Tuesday, Oct. 29, 2019 · 11:32 a.m. · 10m 56s
Finding control algorithm with dynamic simulation - Energy savings and comfort optimization - Web-based HVAC control
Embed
Transcriptions
Note: this content has been automatically generated.
00:00:00
and i'm and i'm happy to a show you what a
00:00:05
about something about uh i would work what we are doing
00:00:09
i'm working since about twelve years i think you with my dies so from
00:00:13
us car i learned that i understood up to now only twenty five percent
00:00:19
so i'm curious to learn about the other seventy five percent of the power of this tool
00:00:27
so
00:00:31
this is my uh these are the points i want a would
00:00:35
like to show you first of all what the spur nor means
00:00:39
and then uh i would like to show you
00:00:41
our approach for a predictive up a control optimisation
00:00:49
and then of course i make to lynch
00:00:51
to how we are using a building performance immolation
00:00:56
and tell you something about the models uh what i'm working with
00:01:02
uh then uh
00:01:05
i will show you another way of a predictive control optimisation if
00:01:10
the building and the eye track systems all those and some conclusions
00:01:16
so first of all what this problem means it's it's it's great
00:01:20
and it's mean it means probably tents so our
00:01:24
company name is also our program so we are uh
00:01:30
we are using waited for cost data and we translates them to uh
00:01:36
control in an optimist way h. rack systems for heating and cooling systems
00:01:44
how are we doing this
00:01:47
we are using we're working with a deterministic uh
00:01:50
approach or you can call it also rule based approach
00:01:56
how does this work
00:01:58
first of all you have a building the kind of building a type of
00:02:03
building and you know uh how does the case uh with the frame conditions
00:02:10
uh for instance the due to the weather conditions
00:02:16
but you know also you have for heating cooling systems so
00:02:19
this is also a control by controllers you know how this works
00:02:26
so you can imagine you are i'm an engineer with a lot of experience in building and then
00:02:31
the cheap and you understand how the system works
00:02:34
more or less if once you understand how it works
00:02:38
then you can make rules rules to option
00:02:42
mice the the control of heating and cooling systems
00:02:48
if you look at the weather forecast this is in few words explained the rule based
00:02:55
approach and you when you see now also
00:02:58
the difference to the model predictive control approach
00:03:02
maybe you uh assists to the very
00:03:05
forced the presentation uh yesterday afternoon three o'clock
00:03:10
they are uh does presentation showed the model predictive control
00:03:14
approach and the uh now you can see the difference
00:03:21
how do we use uh idle eyes or
00:03:25
building performance immolation to to to do this
00:03:28
first of all you have to make and model
00:03:31
of the building and also of the h. rack systems
00:03:36
then you have to model lies the rules but i just mentioned before
00:03:41
you take a typical metallurgical your data and
00:03:45
very important you have to tell the model
00:03:48
your name so what what would you like to achieve with this optimised control
00:03:55
in our uh case it's this minimise energy
00:03:58
consumption total energy consumption and maximise comfort conditions
00:04:06
now how how to what what what does the model uh dust
00:04:13
what kind of of a model to we used to to do this
00:04:18
there are some uh standard uh mm elements in
00:04:22
this level like the building poppy side shading orientation
00:04:28
then some typical sounds what we uh and allies the comfort and energy consumption
00:04:35
of course the the construction of the building user behaviour
00:04:39
h. rex systems this is standard building performance english what
00:04:44
special case that we also model lust
00:04:48
the rules to these optimised pretty if control
00:04:54
uh then you have also the weather forecast data which i take
00:05:00
the typical metallurgical your data and with excel at shift them in time
00:05:05
so i have weather forecast data they a hundred percent correct because it's only time she
00:05:13
and it's with a graphical scripts i can then determine the target
00:05:19
the target parameter so does a minimise then the g. and maximise comfort
00:05:27
some short in the new uh for pros
00:05:31
who have seen uh this already a little bit
00:05:35
so uh those who who know i dice they
00:05:38
recognise very quickly here stand that a plant macro with
00:05:44
a boiler regulation what is no special is
00:05:48
that uh our mark role with the rules
00:05:53
of course i won't tell you now all the tape i about these rules because that's our know how
00:05:59
but uh you see here a little bit how uh it contains
00:06:05
so first of all there are a constant inputs so this means that the weather
00:06:09
for to to with the fork a state that time shift that t. n. y. data
00:06:15
and then the variable input and this variable input
00:06:20
i gave the freedom to to to to
00:06:23
battery this input to look for the optimum result
00:06:30
i'll thought then is an offset and this offset
00:06:33
overwrites the standard set points of paycheck controls it
00:06:39
mm kay bases the graphic postscript to determine
00:06:44
to calculate a at least a target parameter
00:06:49
so u. k. hit the comfort parameter the energy
00:06:52
consumption then you have to enormous these two parameters
00:06:58
you can if you want to if different weights to one or the other parameter
00:07:04
if you would like to take them in a in its anyway so you put two times one
00:07:11
and these gifts at this output parameter which is
00:07:15
the uh determines the behaviour of the parametric around too
00:07:23
so this was in short works how we're using
00:07:27
building performance immolation to uh develop a predictive control systems
00:07:33
which uh works with the central lust h. back control devices
00:07:40
but if you're building a maybe is a big office buildings
00:07:45
with uh it's some why is regulation devices like here for instance
00:07:50
then you can do the same also not only by the centralise the controls also
00:07:56
right that the centralise the controls for instance for this room here which has a somewhere
00:08:02
a temperature sensor and its own regulation the soap assistance
00:08:10
and of course everyone here know step
00:08:13
the simple room temperature regulation is always behind
00:08:18
i'm a real a reaction because uh it it doesn't look ahead
00:08:26
and with building performance immolation you can't very
00:08:29
well also uh anticipates the behaviour of the room
00:08:37
this and this is the
00:08:40
for two days you have to change uh in account the
00:08:44
site shaking because it's very dynamic h. the the solar gains
00:08:50
uh are very dynamic due to the location of the sounds in the building
00:08:57
and a slow to react on this uh dynamic h. u. u. k. and
00:09:03
uh apply uh an office a set point oh right uh to each seven
00:09:13
of course when you have a simple
00:09:15
building without um automatic that uh shading control
00:09:20
for example then the only dynamic in the building is the shade moving uh uh
00:09:28
from this uh from some but if you have oh to
00:09:31
make that a shading controls like a and there's a explain before
00:09:38
then you have another dynamic but this dynamic is known
00:09:41
because you know the algorithm of that the shaking control
00:09:45
so then you have to take in account how the shaking control uh works
00:09:51
to determine a t. optimise the
00:09:55
control algorithm for heating and cooling says
00:10:01
but i'm coming to conclusions so for me
00:10:05
uh extract control systems uh cannot be a independent
00:10:11
uh from the facets design the facet design contains a shading
00:10:17
systems these shape or contains also a variable a solar loads
00:10:25
and the therefore you can to take this into
00:10:30
account with in the uh the eye track control strategies
00:10:36
and for a dynamic similar simulation tools like ida ice
00:10:41
well they uh hello this development of a optimised control algorithms
00:10:48
no ice they they up for questions thank you very much for your time ah
Conference Program
Implementing a quality assurance program for sustainable buildings
Don Neff, LJP Construction Services, Irvine, United States
Oct. 28, 2019 · 4:34 p.m.
317 views
Performance of heat recovery using an Active Double Skin Façade
Michael Bulander & Kishore Patel, Harley Ellis Devereaux, Los Angeles, United States
Oct. 28, 2019 · 4:47 p.m.
468 views
A bio-inspired solution to raise the efficiency of kinetic façades
Steven Ware, Art & Build Architects, Paris, France
Oct. 28, 2019 · 5:36 p.m.
155 views
Concrete curtain walls - Status, benefits and visions
Stephan Giesser, Solidian, Germany
Oct. 29, 2019 · 8:32 a.m.
153 views
Advances in passive and active coatings technology for glass
Rory Back, NSG Pilkington, Lathom, United Kingdom
Oct. 29, 2019 · 9:59 a.m.
194 views
Why shading must be dynamic
Anders Hall, Somfy, Sweden
Oct. 29, 2019 · 10:46 a.m.
745 views
Glazing and shading that make the difference
Oskar Storm, Saint-Gobain Glass, Sweden
Oct. 29, 2019 · 11:01 a.m.
5344 views
Successful façade design by simulation
Sven Moosberger, EQUA Solutions, Switzerland
Oct. 29, 2019 · 11:17 a.m.
181 views
Digitally printed multi-colored PV for building integration
Dirk Hengevoss & Michaela Terwilliger, Resp: FHNW, Switzerland - Hochschule Luzern, Switzerland
Oct. 29, 2019 · 11:20 a.m.
Predictive controls for HVAC systems
Urs Grossenbacher, Pronoó AG, Givisiez, Switzerland
Oct. 29, 2019 · 11:32 a.m.
Holistic, multi-dimensional analysis of shading impacts on an all-glass office building
Jason Kirkpatrick, Interface Engineering, San Francisco, United States
Oct. 29, 2019 · 11:44 a.m.
Advanced daylight modeling of façade systems for energy and comfort analysis
David Geisler-Moroder, Bartenbach GmbH, Aldrans, Austria
Oct. 29, 2019 · 12:16 p.m.
New building glass envelope with Light, Color and Images
Ion Luh, Consullux Lighting Consultants/CEL, Canada
Oct. 29, 2019 · 2:49 p.m.
The Desert Rose - When poesy meets technology
Nassim Saoud, Trimble Consulting, Paris, France
Oct. 29, 2019 · 4:30 p.m.
372 views
