Keywords: thermal

Summary

This demo shows how the data-driven temperature solver and an external temperature vs. time file can be used to drive the temperature of an object in a DIRSIG simulation. In this demo, a simple utility pole with three transformer "cans" is constructed. One of the transformer "cans" is passively heated by the Sun (using the built-in THERM temperature solver) and the remaining two are driven by unique temperature vs. time data files. The simulation generates an LWIR image every hour over the course of a 24 hour period so that the variation in the various objects can be observed.

Details

The data-driven temperature solver allows the user to specify either

  1. A temporally constant temperature, or

  2. A temporally varying temperature.

Since temperature solvers are assigned to materials, it means that all surfaces assigned this material would get assigned the corresponding temperature. The user can distribute material assignments within their geometry to facilitate the spatial granularity that temperatures are assigned at.

Note
 The spatial resolution with which temperatures can be mapped is driven by the granulatity with which materials are assigned. Therefore, this approach is best suited for situations where the material properties and temperature signatures are well correlated to a geometric component.

This demo features two objects (with two unique materials) that are configured with the data-driven temperature solver. The temperature vs. time files for this demo are simple sin2 functions, but a real world application might import this data from an external model that is specialized to predict the temperature of a given target object.

Important Files

This section highlights key files important to the simulation.

Scene Geometry

The geometry for this scene is completely described using the built-in geometry objects via the GLIST file (see the demo.glist file).

  • The tall (12 meter) pole is a single cylinder object.

  • The 3 cross beams are 3 instances of the same box object.

  • The 3 transformer cans are 3 unique cylinder objects (each cylinder needs unique material ID).

Atmosphere Configuration

For this simulation, the Simple atmosphere model was used with the default sky temperature of 280 K.

Material Configuration

The various material configurations can be explored using the Material Database Editor or by hand-editing the DIRSIG .mat file. The scene utilizes 4 unique materials:

  1. Transformer #1 (ID #1, metal, THERM temperature solver)

  2. Transformer #2 (ID #2, data-driven temperature solver)

  3. Transformer #3 (ID #3, data-driven temperature solver)

  4. Wood (ID #10, used for the pole and cross beams, THERM temperature solver)

All the materials in the scene use the same 10% Lambertian reflectance (0.90 emissivity) optical property. The primary difference between all the materials is the temperature solver configuration. The "wood" and transformer #1 (passive) temperature solver configurations were forged from the THERM properties used in other scenes. The transformer #2 and #3 temperature solvers are set to use the data-driven model using an external temperature vs. time ASCII/text file. The input temporal temperature profiles can be found in the materials/temps folder. These profiles were crafted in Excel from simple analytical function. The first profile is a sin2 function with a two day period (a single peak during the day). The second profile is a sin2 function with a one day period (two peaks during the day). Each profile was then exported to a 2-column ASCII/Text file. These two profiles are plotted below:

materials/temps/temps.png
Figure 1. Diurnal temperature profiles for the two data-driven transformers.

Platform Configuration

The imaging platform has a 320 x 240 (QVGA format) LWIR camera with a passband response of 8 - 12 microns. The platform is located on the ground looking up at the utility pole from 30 meters away.

Setup

The simulation can be recreated by loading and running the simulation scenario in the demo.sim file. The output of the simulation is 24 radiance image and 24 truth image files (one file for each hourly capture).

Results

The diurnal animation below was created from the Avg. Temperature truth. The individual frames in the animation use the same gain throughout the day.

demo.gif
Figure 2. Diurnal animation of the scene showing the passive and data-driven temperature changes throughout the day.

The sky is a constant radiance throughout the day due to the Simple atmosphere model configuration, so the brightness does not fluctuate. The utility pole and left transformer (Transformer #1) warm up and cool down in response to the solar loading. The center (Transformer #2) and right (Transformer #3) "cans" modulate in response to the data-driven temperature profiles supplied to the respective data-driven temperature solver.