White Box Technologies brings simulation weather data to the satellite age

14 posts / 0 new
Last post

Hi Joe,

From your perspective, are NREL or any other government/professional bodies making moves/indications to update/refresh the current TMY3 sets to utilize satellite-derived solar radiation? Seems like a no-brainer for our industry, but is there a counter-argument? It seems likely, but has there been rigorous comparisons of satellite-derived solar radiation against measured values and/or our ?present-day? solar models used to derive solar radiation information for building energy simulation?

Also, I?m trying to understand and correctly characterize the impact of this development in simple terms my brain can follow. Is it fair to say:

The solar models used in developing weather files for building energy simulations to-date in our industry (including all/most industry-standard TMY weather sets), because they have been using solar radiation derived from (evolving) solar models, have not accounted for the likes of local climate cloud cover / smog / dust? Seattle (~47?N) has perhaps been seeing as much sunlight through the winter as Paris (~48?N)?

Does satellite-derived solar radiation address some or all of these local climate issues (cloud cover, smog, dust) affecting direct/indirect solar radiation?

Thanks sincerely for all your teaching Joe,

~Nick

[cid:image001.png at 01D383A3.92EEFC80]
Nick Caton, P.E., BEMP
Senior Energy Engineer
Regional Energy Engineering Manager
Energy and Sustainability Services
Schneider Electric

D 913.564.6361
M 785.410.3317
F 913.564.6380
E nicholas.caton at schneider-electric.com

15200 Santa Fe Trail Drive
Suite 204
Lenexa, KS 66219
United States

[cid:image002.png at 01D383A3.92EEFC80]

Nicholas.Caton at schneider-electric.com's picture
Joined: 2016-02-26
Reputation: 0

Nick,

This is an interesting topic that has evolved in an unexpected way.? Now that NREL has
mastered the technology of satellite-derived solar radiation, they have regarded weather
station data as an impediment, since there's no way to get such measured data to match the
5-kilometer grid of the satellite-derived solar.? Therefore, they have abandoned the
weather station data (which was was used in all the TMYs to date) and gone instead to
Reanalysis Data from NOAA's MERRA, which is running a climate forecasting model in
retrospective mode. I don't have time now to discuss reanalysis, except to say that from
what I've seen the results are decidedly "iffy".? Almost two years ago, I managed to get a
Work Statement through ASHRAE? for someone to take a good look at reanalysis data.

So, what does this mean for your question??? NREL's National Solar Radiation Data Base
(NSRDB) can now get you the hourly time series or a TMY for more than a million grid
points over the US, all with satellite-derived solar radiation but? MERRA results for the
other climate parameters.? I've told NREL several years ago that for the building
simulation industry it would be much better to merge the? satellite-derived solar with
actual station data that give good accurate coverage of urban areas where buildings are
located. NREL's response was that they'd be happy to do this, but someone has to pay them
for the work.

In response to your hypothesis below, I think you're being too disparaging of the previous
modeling efforts while raising too high your expectations of satellite solar.? It's not
that the previous models failed to account for local climate conditions, but that they
lacked good data to drive them.? For example, all models included terms for cloud cover
and clearness, or for the more detailed physical models arcane parameters like aerosol
optical depth, preciptable moisture, etc.,? but how available are the input data and how
reliable are they?? The advantages of satellite-derived solar are that they provide a
comprehensive and objective view of the cloud conditions, which combined with satellite
measurements of the atmospheric conditions and? improved modeling, results in? accuracies
that previous modeling efforts can not attain.? As far as discerning localized effects of
smog and dust in urban areas, that would still depend on whether there's sufficient
monitoring at that spatial and time scale to detect the differences.?? What I mean is that
it's one thing to observe that in general urban locations have more smog and particulates
than rural locations, but it's something else to quantify the resultant differences in
solar radiation over time and distance.

I'd like to take the opportunity here to step back and comment on the status of weather
data for the building energy community, My interactions with NREL has brought the
realization that we have been piggy-backing on the efforts of others outside our community
for our weather data.? I don't intend to pick on the NREL Solar Program, several of whom I
consider friends and colleagues, but their target client is the solar power industry.?
Since solar power arrays can be installed anywhere, preferably in rural uninhabited
locations, it makes sense to go to satellite-derived solar.? It's also clear that to serve
that industry, NREL would focus its efforts on getting the best solar values, while all
the other climatic parameters, like temperature, humidity, wind speed, etc., are
secondary, which may be why getting them from MERRA is a satisfactory choice.? The focus
on solar is also evident in the weighting used by NREL to develop the TMYs, with 50%
weight placed on the 2 solar and 50% on the remaining 8 non-solar parameters.

For the building energy community, or priorities are somewhat different. Since 99% of
buildings are located in urban locations, we should focus much more on climate in urban
areas.? Luckily, that's also where the great majority of existing weather are located,
which is why I'm resistant to throwing out measured weather data and replacing them with
synthetic data, no matter how much they've been "seeded" with real data. As for the
weighting of climate parameters in selecting the typical months,? why not use building
energy simulations and weight them by the distribution of heating and cooling loads?

Joe

Joe Huang
White Box Technologies, Inc.
346 Rheem Blvd., Suite 205A
Moraga CA 94556
yjhuang at whiteboxtechnologies.com
http://weather.whiteboxtechnologies.com for simulation-ready weather data
(o) (925)388-0265
(c) (510)928-2683
"building energy simulations at your fingertips"

Joe Huang's picture
Offline
Joined: 2011-09-30
Reputation: 206

Nick,

This is an interesting topic that has evolved in an unexpected way.? Now that NREL has
mastered the technology of satellite-derived solar radiation, they have regarded weather
station data as an impediment, since there's no way to get such measured data to match the
5-kilometer grid of the satellite-derived solar.? Therefore, they have abandoned the
weather station data (which was was used in all the TMYs to date) and gone instead to
Reanalysis Data from NOAA's MERRA, which is running a climate forecasting model in
retrospective mode. I don't have time now to discuss reanalysis, except to say that from
what I've seen the results are decidedly "iffy".? Almost two years ago, I managed to get a
Work Statement through ASHRAE? for someone to take a good look at reanalysis data.

So, what does this mean for your question??? NREL's National Solar Radiation Data Base
(NSRDB) can now get you the hourly time series or a TMY for more than a million grid
points over the US, all with satellite-derived solar radiation but? MERRA results for the
other climate parameters.? I've told NREL several years ago that for the building
simulation industry it would be much better to merge the? satellite-derived solar with
actual station data that give good accurate coverage of urban areas where buildings are
located. NREL's response was that they'd be happy to do this, but someone has to pay them
for the work.

In response to your hypothesis below, I think you're being too disparaging of the previous
modeling efforts while raising too high your expectations of satellite solar.? It's not
that the previous models failed to account for local climate conditions, but that they
lacked good data to drive them.? For example, all models included terms for cloud cover
and clearness, or for the more detailed physical models arcane parameters like aerosol
optical depth, preciptable moisture, etc.,? but how available are the input data and how
reliable are they?? The advantages of satellite-derived solar are that they provide a
comprehensive and objective view of the cloud conditions, which combined with satellite
measurements of the atmospheric conditions and? improved modeling, results in? accuracies
that previous modeling efforts can not attain.? As far as discerning localized effects of
smog and dust in urban areas, that would still depend on whether there's sufficient
monitoring at that spatial and time scale to detect the differences.?? What I mean is that
it's one thing to observe that in general urban locations have more smog and particulates
than rural locations, but it's something else to quantify the resultant differences in
solar radiation over time and distance.

I'd like to take the opportunity here to step back and comment on the status of weather
data for the building energy community, My interactions with NREL has brought the
realization that we have been piggy-backing on the efforts of others outside our community
for our weather data.? I don't intend to pick on the NREL Solar Program, several of whom I
consider friends and colleagues, but their target client is the solar power industry.?
Since solar power arrays can be installed anywhere, preferably in rural uninhabited
locations, it makes sense to go to satellite-derived solar.? It's also clear that to serve
that industry, NREL would focus its efforts on getting the best solar values, while all
the other climatic parameters, like temperature, humidity, wind speed, etc., are
secondary, which may be why getting them from MERRA is a satisfactory choice.? The focus
on solar is also evident in the weighting used by NREL to develop the TMYs, with 50%
weight placed on the 2 solar and 50% on the remaining 8 non-solar parameters.

For the building energy community, or priorities are somewhat different. Since 99% of
buildings are located in urban locations, we should focus much more on climate in urban
areas.? Luckily, that's also where the great majority of existing weather are located,
which is why I'm resistant to throwing out measured weather data and replacing them with
synthetic data, no matter how much they've been "seeded" with real data. As for the
weighting of climate parameters in selecting the typical months,? why not use building
energy simulations and weight them by the distribution of heating and cooling loads?

Joe

Joe Huang
White Box Technologies, Inc.
346 Rheem Blvd., Suite 205A
Moraga CA 94556
yjhuang at whiteboxtechnologies.com
http://weather.whiteboxtechnologies.com for simulation-ready weather data
(o) (925)388-0265
(c) (510)928-2683
"building energy simulations at your fingertips"

Joe Huang's picture
Offline
Joined: 2011-09-30
Reputation: 206

I have a question ? what happened to the rain and snow flags that were in the old CWEC file but are missing in the latest versions. Many snow melting systems have a snow/rain sensor in the slab used to trigger the system on.

Christopher R. Jones, P.Eng.
Technical Specialist
Sustainability & Energy

[cid:image001.png at 01D38893.836DAF50]
T +1 416-644-0252

2300 Yonge Street, Suite 2300
Toronto, ON M4P 1E4 Canada

wsp.com

Please consider the environment before printing...

Jones, Christopher2's picture
Joined: 2017-10-12
Reputation: 0

The rain and snow flags are still available in the raw weather station data, but haven't
been included in the processed weather files since the 1990's because? everyone in this
hemisphere at least have adopted the TMY2/TMY3 formats, in content? if not the literal
file format, developed by NREL, which does not include these two flags. I was also under
the impression that none of the simulation programs, such as DOE-2, do not use these two
flags anyway.?? If these flags are of use, it would make sense to include them, which
would not require a change in the DOE-2 *.BINM format, but would require a change to the
EnergyPlus *.epw format.

Joe

Joe Huang
White Box Technologies, Inc.
346 Rheem Blvd., Suite 205A
Moraga CA 94556
yjhuang at whiteboxtechnologies.com
http://weather.whiteboxtechnologies.com for simulation-ready weather data
(o) (925)388-0265
(c) (510)928-2683
"building energy simulations at your fingertips"

Joe Huang's picture
Offline
Joined: 2011-09-30
Reputation: 206

I have a question ? what happened to the rain and snow flags that were in the old CWEC file but are missing in the latest versions. Many snow melting systems have a snow/rain sensor in the slab used to trigger the system on.

Christopher R. Jones, P.Eng.
Technical Specialist
Sustainability & Energy

[cid:image001.png at 01D38893.836DAF50]
T +1 416-644-0252

2300 Yonge Street, Suite 2300
Toronto, ON M4P 1E4 Canada

wsp.com

Please consider the environment before printing...

Jones, Christopher2's picture
Joined: 2017-10-12
Reputation: 0

The rain and snow flags are still available in the raw weather station data, but haven't
been included in the processed weather files since the 1990's because? everyone in this
hemisphere at least have adopted the TMY2/TMY3 formats, in content? if not the literal
file format, developed by NREL, which does not include these two flags. I was also under
the impression that none of the simulation programs, such as DOE-2, do not use these two
flags anyway.?? If these flags are of use, it would make sense to include them, which
would not require a change in the DOE-2 *.BINM format, but would require a change to the
EnergyPlus *.epw format.

Joe

Joe Huang
White Box Technologies, Inc.
346 Rheem Blvd., Suite 205A
Moraga CA 94556
yjhuang at whiteboxtechnologies.com
http://weather.whiteboxtechnologies.com for simulation-ready weather data
(o) (925)388-0265
(c) (510)928-2683
"building energy simulations at your fingertips"

Joe Huang's picture
Offline
Joined: 2011-09-30
Reputation: 206

Actually, EnergyPlus (and the epw format) uses
coded flags in the 9 digit "Present Weather
Codes" to determine and set Rain/Snow. Now that
some stations include precipitation (albeit
liquid) EnergyPlus sets the rain flag when there
is adequate precipitation currently.

I believe the updated TMY3 format went away from
the 9 digit present weather codes, so EnergyPlus
would rely on the "preciitation" field in the epw file.

It would not be impossible to bring those back to
the CWEC data; however, probably not as individual flags.

Linda

Linda Lawrie2's picture
Offline
Joined: 2011-10-02
Reputation: 0

Thanks Joe,
I believe you are correct in saying that the simulation programs don?t make use of the rain/snow flags. We only use them in our spreadsheet calculations for determining snow melting energy. We work on a number of commuter train/bus stations with very large areas with snow melting systems. Most of the control sequences include slab sensors to detect snow or rain.

Christopher R. Jones, P.Eng.
Technical Specialist
Sustainability & Energy

[cid:image001.png at 01D38935.300285A0]
T +1 416-644-0252

2300 Yonge Street, Suite 2300
Toronto, ON M4P 1E4 Canada

wsp.com

Please consider the environment before printing...

Jones, Christopher2's picture
Joined: 2017-10-12
Reputation: 0

I?ve grown sensitive to ensuring photovoltaic array models (using SAM) utilize weather data that incorporates snow. Based on my experiences looking at array designs in Illinois / Indiana over the last year, snow accumulation can certainly be the largest source of direct production losses in many climates, so it can be a big thing to miss if it isn?t in the weather you feed in.

I?m not aware of how any simulation tools/platforms outside of SAM however that directly utilize precipitation data.

~Nick

[cid:image002.png at 01D38935.44DC5A00]
Nick Caton, P.E., BEMP
Senior Energy Engineer
Regional Energy Engineering Manager
Energy and Sustainability Services
Schneider Electric

D 913.564.6361
M 785.410.3317
F 913.564.6380
E nicholas.caton at schneider-electric.com

15200 Santa Fe Trail Drive
Suite 204
Lenexa, KS 66219
United States

[cid:image003.png at 01D38935.44DC5A00]

Nicholas.Caton at schneider-electric.com's picture
Joined: 2016-02-26
Reputation: 0

Thanks Joe,
I believe you are correct in saying that the simulation programs don?t make use of the rain/snow flags. We only use them in our spreadsheet calculations for determining snow melting energy. We work on a number of commuter train/bus stations with very large areas with snow melting systems. Most of the control sequences include slab sensors to detect snow or rain.

Christopher R. Jones, P.Eng.
Technical Specialist
Sustainability & Energy

[cid:image001.png at 01D38935.300285A0]
T +1 416-644-0252

2300 Yonge Street, Suite 2300
Toronto, ON M4P 1E4 Canada

wsp.com

Please consider the environment before printing...

Jones, Christopher2's picture
Joined: 2017-10-12
Reputation: 0

I?ve grown sensitive to ensuring photovoltaic array models (using SAM) utilize weather data that incorporates snow. Based on my experiences looking at array designs in Illinois / Indiana over the last year, snow accumulation can certainly be the largest source of direct production losses in many climates, so it can be a big thing to miss if it isn?t in the weather you feed in.

I?m not aware of how any simulation tools/platforms outside of SAM however that directly utilize precipitation data.

~Nick

[cid:image002.png at 01D38935.44DC5A00]
Nick Caton, P.E., BEMP
Senior Energy Engineer
Regional Energy Engineering Manager
Energy and Sustainability Services
Schneider Electric

D 913.564.6361
M 785.410.3317
F 913.564.6380
E nicholas.caton at schneider-electric.com

15200 Santa Fe Trail Drive
Suite 204
Lenexa, KS 66219
United States

[cid:image003.png at 01D38935.44DC5A00]

Nicholas.Caton at schneider-electric.com's picture
Joined: 2016-02-26
Reputation: 0

The 9-digit Present Weather Code (PW) was used only in North America before the turn of
the century an long since replaced by a METAR 2-digit PW code now used in? weather reports
around the world.? So, if the epw? format leaves space for this old coding, it's
functionally dead space and probably filled with 999999999 in all files except the old
TMY2s and CWECs. The Liquid Precipitation (LP) is very different because it gives the
amount of precipitation and time over which it occurred, typically in multiples of 6 up to
24 hours, not whether or not there is rain during that hour.?? This makes LP reports
basically unusable in simulations, which is why I spent several weeks back in 2015 to
convert the LP in the TMY3s to hourly records.? So, if EnergyPlus is relying on the LP
field to detect the presence of rain, make sure that the LP has undergone such a data
transformation.? And what about snow flag, which seems to be of greater interest ?

In my opinion, if the rain and snow flags are useful, we should take them from the current
2-digit METAR PWC.? I've already doing that? starting with the IWEC2 weather files, where
I added an extra field for PW? to the end of the TMY3 format,? although I never made a big
fuss about the format change.? PW appears on all the weather files? that I produce in the
*.FIN4 format, but it's lost in the conversion to epw nor have I bothered to? use them to
set the ISNOW and IRAIN? flags in the DOE-2 bin files.? To do so is very easy, since in
the METAR PWC all? instances of rain have the same initial digit, as do all instances of
snow.

Joe

Joe Huang
White Box Technologies, Inc.
346 Rheem Blvd., Suite 205A
Moraga CA 94556
yjhuang at whiteboxtechnologies.com
http://weather.whiteboxtechnologies.com for simulation-ready weather data
(o) (925)388-0265
(c) (510)928-2683
"building energy simulations at your fingertips"

Joe Huang's picture
Offline
Joined: 2011-09-30
Reputation: 206
Aaron Powers2's picture
Offline
Joined: 2011-09-30
Reputation: 0