Its all about Conserving Energy How predictive rendering can avoid - - PowerPoint PPT Presentation

Ian Williams, Director of Applied Engineering, NVIDIA

Its all about Conserving Energy

How predictive rendering can avoid creating an accidental Death Ray

2

Agenda

The Problem A Quick Physics Recap Simulating the Death Ray in Iray Deriving Temperature from Irradiance Comparing Predictions with Measurement Conclusions

3

The Problem

August 29th 2013

Martin Lindsay parked his Jaguar car on Eastcheap, London around noon 2 hours later he returned to find parts of the car had melted

4

The Scale of the Problem

“On Tuesday afternoon, I was sent out to see if I could fry an egg in the heat, a task that I presumed was impossible on an overcast September day. But, not only was it possible, I had to run out of the death ray that was slowly cooking my egg, because the thinning hairs on my head started to catch fire. ” “On Monday, the air temperature in the concentrated beam, reached 69.8C, which in old money is 158F. To put that in context, the world’s hottest temperature was recorded in Death Valley at 56.7C (134F) over a century ago.”

5

Death Ray’s Apparently are Fairly Common

6

Agenda

A Quick Physics Recap

7

Electromagnetic Radiation

Electromagnetic Radiation consists of waves that are synchronized oscillations of electric and magnetic fields that propagate at the speed of light Produced whenever charged particles are accelerated, and these waves can subsequently interact with any charged particles EM waves carry energy, momentum and angular momentum away from their source particle and can impart those quantities to matter with which they interact Electromagnetic spectrum: radio waves, microwaves, infrared radiation, visible light, ultraviolet radiation, X-rays and gamma rays

A Quick Recap

8

How Does Sunlight Heat Things Up?

Radiation is both absorbed and emitted Convection Conduction

Three Main Modes of Energy Transfer

9

Governing Laws

The law of conservation of energy states:

That the total energy of an isolated system is constant; energy can be transformed from one form to another, but cannot be created or destroyed

The first law of thermodynamics is a version of the law of conservation of energy, adapted for thermodynamic systems, and states:

That the change in the internal energy of a closed system is equal to the amount of heat supplied to the system, minus the amount of work done by the system on its surroundings

10

Quantifying Energy Transfer in the Electro Magnetic Spectrum

Radiant Flux per unit area received by a Surface Units = watts per square metre W/m2 Hemispherical quantity

Irradiance

11

Solar Irradiance and Atmospheric Absorbtion

The electromagnetic radiation emitted by the sun incident on the top of the atmosphere

50% in region beyond visible light 40% in visible light region 10% in ultraviolet region

On a clear day approximately 50% of solar energy absorbed by atmosphere

Irradiance

12

Agenda

Simulating the Death Ray in Iray

13

Simulating the Physics of the Death Ray

Walkie Talkie Building and Environment were extensively modelled in detail using AutoDesk 3D Studio Max Used Iray Renderer to:

Simulate time of day and year using calibrated Sunlight model Calculate Irradiance incident on specific surface

Iray

14

The Model

Movie 20 Fenchurch Street and Surroundings, London

15

Time of Day Analysis

Movie 29 August 2013

16

Predicted Irradiance from Iray

Movie 29 August 2013

17

The Energy Contained in the Death Ray

Direct Sunlight on a clear day has an Irradiance of ~110K Lux or ~1.1 KW/m2 Predicted Irradiance at street level outside the Death Ray was ~98K Lux or ~1.1 KW/m2 Predicted Irradiance at street level inside the Death Ray was ~700K Lux or ~7.5 KW/m2 =6x more powerful than the Sun!

18

It Could Have Been Worse...

Slight modifications to the building geometry, in the form of between 5 to 10 degrees more curvature, increased Irradiance from: ~700K Lux or ~7.5 KW/m2 To ~3M Lux or ~32.3 KW/m2 =30x more powerful than the Sun!!!

19

Agenda

Deriving Temperature from Irradiance

20

Calculating Temperature from Irradiance

Solar Heat Transfer to a Surface

21

Temperature of the Surface

Assuming system is steady state with no phase changes, then from 1st law of thermodynamics the energy balance equation is:

Energy Balance

Ee . α = ε . σ [ T

surf-air 4 - T air-amb 4 ] + hsurf-air [T surf-air – T air-amb] + k . [T surf-air – T surf-other] / s

Absorbed Energy Re-Radiated Energy Energy transfer through Convection Energy transfer Through Conduction

Surf -> Atmosphere Atmosphere

• > Surf

Ee = Irradiance (Wm-2) α = Absorptivity ε = Emissivity σ = Stefan-Boltzman constant = 5.670373x10-8 (Wm-2K-4) T

surf-air = Temperature at air surface boundary (K)

T

air-amb = Ambient air temperature (K)

T

surf-other = Temperature at other side of surface (K)

hsurf-air = Heat transfer coefficient between surface and air (Wm-2K-1) k = Thermal Conductivity of surface material (Wm-1K-1) s = Surface material thickness

22

Simplifying Things

Assume worst case scenario where energy transfer/cooling through conduction is negligible Assume radiation from atmosphere back to surface can be considered negligible Energy Transfer Equation be re-written: Which equals: Which is a quartic of the form:

Ee . α - ε . σ T

surf-air 4 - hsurf-air [T surf-air – T air-amb] = 0

ε . σ . T

surf-air 4 + hsurf-air .T surf-air – [ Ee . α + hsurf-air .T air-amb] = 0

a . x4 + b . x3 + c . X2 + d .x – e = 0 a = ε . σ d = hsurf-air e = Ee . α + hsurf-air .T

air-amb

23

General Solution To A Quartic Equation

24

In Our Case…

b = c = 0

25

So The Formula Becomes…

Δ0 = 12ae Δ1 = 27ad2 Q = [ (Δ1 + (Δ1

2 - 4 Δ0 3)1/2 )/2]1/3

S = ½ [ (1/3a) * (Q + Δ0/Q) ]1/2 q = d/a a = ε . σ d = hsurf-air e = Ee . α + hsurf-air .T

air-amb + ε . σ .T air-amb 4

T

surf-air = X1 or X2 or X3 or X4

Depending on which root is a +ve real number

Where:

x1,2 = -S ± ½( -4S2 + q/S )1/2 x3,4 = S ± ½( -4S2 – q/S )1/2

26

Assumed Values…

Ee = Irradiance (Wm-2) P(W) = Ev(lx) × A(m2) / η(lm/W) , where η = Lumic Efficacy = 0.93 for sunlight α = Absorptivity ε = Emissivity σ = Stefan-Boltzman constant = 5.670373x10-8 Wm-2K-4 T

surf-air = Temperature at air surface boundary (K)

T

air-amb = Ambient air temperature (K) = 300K

hsurf-air = Heat transfer coefficient between surface and air (Wm-2K-1) = 10.45 –v + 10v1/2 , where v is velocity of object in air (source engineering toolbox) = 10.45 Wm-2K-1 for static object

Absorptivity Emissivity Asphalt 0.93 0.93 Polished Aluminum 0.09 0.09 Magnesium Oxide Paint 0.09 0.09 Chromium 0.2 0.2 Sherwin Williams White Paint 0.89 0.87 Concrete 0.6 0.88 Red Brick 0.94 0.93 Polythene Black Plastic 0.94 0.92 Black Paint 0.94 0.94 Human Skin 0.97 0.97

(source engineering toolbox)

27

Calculated Temperatures

Ambient Temperature = 27C, 80F

NORMAL SUNLIGHT INSIDE DEATH RAY Asphalt

82C, 179F 276C, 530F

Asphalt Light Breeze

57C, 135F 208C, 407F

Black Plastic

82C, 179F 277C, 531F

Black Plastic Light Breeze

58C, 136F 209C, 408F

Chrome

44C, 113F 146C, 295F

Chrome Light Breeze

34C, 94F 82C, 180F

28

Melting Point of Plastics

Material Degrees(F) Material Degrees(F) Acetal (CoPo) 400 PBT 500 Acetal (HoPo) 425 PCT 580 Acrylic 425 Peek 720 Acrylic (Mod) 500 PET 540 ABS (MedImp) 400 Polycarbonate 550 ABS (HiImpFR) 420 Polyetherimide 700 CelAcetate 385 Polyethylene (LD) 325 CelButyrate 350 Polyethylene (HD) 400 CelPropionate 350 Polypropylene 350 EVA 350 Polystyrene (GP) 350 LCP 500 Polystyrene (MI) 380 Nylon (6) 500 Polystyrene (HI) 390 Nylon (6/6) 525 Polysulfone 700 Polyamide-imide 650 PPO 575 Polyarylate 700 PVC (Rig/Flex) 350/325 TFE 600

Source: http://plastictroubleshooter.com/ThePlasticTroubleshooter/melt_temps.htm

29

Observations on Calculated Temperatures

Some temperatures seem believable, eg:

Chrome in Normal Sunlight – 44C, 134F Ashpalt in Light Breeze – 57C, 135F

Some temperatures too high, eg:

Asphalt in the Death Ray – 276C, 530F Asphalt in a Light Breeze - 208C, 407F

Conclusion energy transfer model is too simplistic As energy transfer from Irradiance increases

Reality will be greater dissipation through Convection and Conduction

30

Agenda

Comparing Predictions with Measurement

31

Calibration Scene

32

Calibration Scene in Iray

33

Physical Measurements

34

Comparing Measured and Calculated Temperature

Ambient Temperature = 71F, Windspeed ~0 Knots

CALCULATED FROM IRRADIANCE MEASURED 3-11-2015 1:05 PM MEASURED 3-11-2015 1:51 PM* 1 Asphalt (Horizontal)

122F 106F 87F

2 Concrete (Horizontal)

108F 98F 87.5F

3 Concrete (Vertical, ~south)

116F 104F 89F

4 White Painted Steel (Vertical, ~south)

130F 86.5F 76F

5 White Painted Steel (Vertical, ~east)

66F 78.5F 74.5F

6 Black Painted Steel (Vertical, ~south)

131F 99F 80F

7 Black Painted Steel (Vertical, ~east)

114F 92.5F 80F

* Full Cloud Cover for > 30 mins

35

Observations Between Measured and Calculated Temperature

CALCULATED FROM IRRADIANCE MEASURED 3-11-2015 1:05 PM MEASURED 3-11-2015 1:51 PM* Asphalt (Horizontal)

122F 106F 87F

Concrete (Horizontal)

108F 98F 87.5F

Concrete (Vertical, ~south)

116F 104F 89F

White Painted Steel (Vertical, ~south)

130F 86.5F 76F

White Painted Steel (Vertical, ~east)

66F 78.5F 74.5F

Black Painted Steel (Vertical, ~south)

131F 99F 80F

Black Painted Steel (Vertical, ~east)

114F 92.5F 80F

* Full Cloud Cover for > 30 mins

Significant heat loss after full sunlight is obscured by cloud

• Up to 20% temperature drop

Clearly significant heat transfer occurring through:

• Radiation
• Convection
• Conduction

36

Observations Between Measured and Calculated Temperature (cont.)

CALCULATED FROM IRRADIANCE MEASURED 3-11-2015 1:05 PM MEASURED 3-11-2015 1:51 PM* Asphalt (Horizontal)

122F 106F 87F

Concrete (Horizontal)

108F 98F 87.5F

Concrete (Vertical, ~south)

116F 104F 89F

White Painted Steel (Vertical, ~south)

130F 86.5F 76F

White Painted Steel (Vertical, ~east)

66F 78.5F 74.5F

Black Painted Steel (Vertical, ~south)

131F 99F 80F

Black Painted Steel (Vertical, ~east)

114F 92.5F 80F

* Full Cloud Cover for > 30 mins

Results are expected to be higher since conduction of heat away from the surface is being ignored by the model These are large uniform surfaces so local heat transfer effects will not occur Over prediction is approximately ~9-13%

37

Observations Between Measured and Calculated Temperature (cont.)

CALCULATED FROM IRRADIANCE MEASURED 3-11-2015 1:05 PM MEASURED 3-11-2015 1:51 PM* Asphalt (Horizontal)

122F 106F 87F

Concrete (Horizontal)

108F 98F 87.5F

Concrete (Vertical, ~south)

116F 104F 89F

White Painted Steel (Vertical, ~south)

130F 86.5F 76F

White Painted Steel (Vertical, ~east)

66F 78.5F 74.5F

Black Painted Steel (Vertical, ~south)

131F 99F 80F

Black Painted Steel (Vertical, ~east)

114F 92.5F 80F

* Full Cloud Cover for > 30 mins

These surfaces are in very close proximity so local heat transfer effects will occur and the model assumes steady-state energy transfer at a single surface. It doesn’t, for example, model heat flow from one surface to another when both are subject to different Irradiance levels. The materials for surfaces 4 thru 6 are also highly conductive (steel) and very thin (surface 4 & 5 < ⅛”, surface 5 & 6 < ¼”) so the impact

• f conduction will be increased.

38

Observations Between Measured and Calculated Temperature (cont.)

CALCULATED FROM IRRADIANCE MEASURED 3-11-2015 1:05 PM MEASURED 3-11-2015 1:51 PM* Asphalt (Horizontal)

122F 106F 87F

Concrete (Horizontal)

108F 98F 87.5F

Concrete (Vertical, ~south)

116F 104F 89F

White Painted Steel (Vertical, ~south)

130F 86.5F 76F

White Painted Steel (Vertical, ~east)

66F 78.5F 74.5F

Black Painted Steel (Vertical, ~south)

131F 99F 80F

Black Painted Steel (Vertical, ~east)

114F 92.5F 80F

* Full Cloud Cover for > 30 mins

The above inaccuracies with the model are almost certainly causing this under predicition

39

Summarizing Temperature Derived from Irradiance

Temperature could be calculated from Irradiance in certain limited situations

Large uniform surfaces with consistent, steady state exposure to sunlight

Most real-world situations extremely complex

Assumptions in energy transfer model easily break down

For more accuracy heat transfer needs to be modelled with significant detail

Heat transfer through/between solids Multiple surfaces exposed to different Irradiance levels Etc etc

40

Agenda

Conclusions

41

Overall Conclusions

Miscalculating design impact at either design or engineering stage can lead to significant and expensive consequences Knowledge is Power - ability to understand full implications of design choices critical

• Issues may not be observable in required validation points

Interactive rendering combined with calibrated sun model within in Iray facilitated identifying the Death Ray problem literally in minutes Irradiance prediction within Iray is an extremely good indicator of potential heat problems Deriving temperature from predicted Irradiance extremely complex challenge

• requires significant research, investigation and calibration

Thank You!