Office: Doubts about facade area calculation

[marcusfuchs]

What is the problem?

In the office archetype, the areas of outer walls and windows are calculated based on self._est_facade_area, which we get from self._est_facade_area = self._est_outer_wall_area + self._est_win_area. This again uses estimated values:

self._est_outer_wall_area = self.est_factor_wall_area * \
       type_bldg_area ** self.est_exponent_wall
self._est_win_area = self.est_factor_win_area * \
       type_bldg_area ** self.est_exponent_win

I will use an example building with the following setup:

number_of_floors=5,
height_of_floors=3.5,
net_leased_area=5000,
window_layout=2,
office_layout=3,

With office_layout=3 we expect a quadratic building footprint with self._est_length = self._est_width = 33.91. I would assume the facade area of this building (all four sides) to be approximately 33.91 * 4 * 3.5 * 5 = 2373.7 m². From TEASER I get self._est_facade_area = 2736.028092669177, which is not exactly the same, but in the same range.

Things look more dramatic if I switch to office_layout=1. Here, my footprint is defined by

self._est_length = 88.46153846153847
self._est_width = 13.0

For this building, I expect a facade area for all 4 sides of approx. (2*88.46 + 2*13) * 3.5 * 5 = 3551.1 m². Yet, TEASER returns self._est_facade_area = 2736.028092669177

As a result, all window and outer wall areas are significantly smaller than I would expect them to be. I'd be grateful if anyone could help to clarify whether the problem is with my expectations or with the TEASER code?

[mlauster]

@marcusfuchs, I fear it's your expectation. The envelope areas are not estimated with the building's height. It's an empirical function depending on the net floor area. Thus, deviations are likely.

Von meinem Samsung Galaxy Smartphone gesendet.

-------- Ursprüngliche Nachricht -------- Von: marcusfuchs notifications@github.com Datum: 06.03.19 16:34 (GMT+01:00) An: RWTH-EBC/TEASER TEASER@noreply.github.com Cc: Subscribed subscribed@noreply.github.com Betreff: [RWTH-EBC/TEASER] Office: Doubts about facade area calculation (#569)

What is the problem?

In the office archetype, the areas of outer walls and windows are calculated based on self._est_facade_area, which we get from self._est_facade_area = self._est_outer_wall_area + self._est_win_area. This again uses estimated values:

self._est_outer_wall_area = self.est_factor_wall_area * \

   type_bldg_area ** self.est_exponent_wall

self._est_win_area = self.est_factor_win_area * \

   type_bldg_area ** self.est_exponent_win

I will use an example building with the following setup:

number_of_floors=5,

height_of_floors=3.5,

net_leased_area=5000,

window_layout=2,

office_layout=3,

With office office_layout=3 we expect a quadratic building footprint with self._est_length = self._est_width = 33.91. I would assume the facade area of this building (all four sides) to be approximately 33.91 4 3.5 * 5 = 2373.7 m² . From TEASER I get self._est_facade_area = 2736.028092669177, which is not exactly the same, but in the same range.

Things look more dramatic if I switch to office_layout=1. Here, my footprint is defined by

self._est_length = 88.46153846153847

self._est_width = 13.0

For this building, I expect a facade area for all 4 sides of approx. (288.46 + 213) 3.5 5 = 3551.1 m². Yet, TEASER returns self._est_facade_area = 2736.028092669177

As a result, all window and outer wall areas are significantly smaller than I would expect them to be. I'd be grateful if anyone could help to clarify whether the problem is with my expectations or with the TEASER code?

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHubhttps://github.com/RWTH-EBC/TEASER/issues/569, or mute the threadhttps://github.com/notifications/unsubscribe-auth/AGkskt8s3NEApYlM32cDSagB3yCZLb0Aks5vT-ANgaJpZM4bhK1A.

[marcusfuchs]

@mlauster Thanks a lot for the fast response! And yes, it's perfectly fine if it's just my expectations here. Just if you can spare a few more minutes on this, could you explain the reasoning behind using the empirical function instead of calculating the actual facade areas? I'm just naively assuming that as all the necessary information is there, we could simply calculate the facade area as described above? But I also assume there's a reason behind using the empirical function and that this approach also has advantages?

[mlauster]

I can't recall how estimated length and width are calculated and for what they are exactly used. But given you rely on the actual input, the net floor area, using a correlation to this is the chosen option. That's the procedure described in the underlying statistical publication

[marcusfuchs]

From what I see in the code, the est_length and est_width are calculated based on the footprint area (net_leased_area / number_of_floors) with both being equal for office_layout = 3 or one being set to a fixed value and the other calculated accordingly. They are used to assign the correct share of the total facade area to each orientation. From what I see, this is all working correctly.

My doubts are just that the total facade area before that is estimated. In contrast, all the information is there to calculate the "correct" total facade area. Therefore, if we do not know of any advantages to the estimation, I would suggest to not use the estimation and instead calculate the total facade area from the input data taking the height of the building into account. This can be done in one line of code and would, in my opinion, greatly improve the results we get.