Testing and Simulation

Testing and Simulation

Testing window coverings at FLEXLAB
Testing window coverings at FLEXLAB®

The thermal performance of in-plane products can, in general, be simulated similar to sealed (insulated) glazing with modifications to account for long-wave (IR) radiant transmission, gas flow across attachment layers, and shape factors affecting convection over the surface.

LBNL has developed several attachment models including: cellular shades, vertical louvered blinds and perforated screens. In addition, CGDB, the database containing complex product optical performance data, has been completed and publicly released. In order to have confidence in the newly developed attachment models, they must be validated through extensive testing and detailed CFD simulation. This paper outlines a review and validation of the ISO 15099 center-of-glass (COG) heat transfer correlations for naturally ventilated cavities through measurement and simulation.

Three basic methods are used to determine heat flow though window + attachment configurations; physical testing through steady state measurements, two-dimensional finite element conjugate heat transfer and computational fluid dynamics (FEM) simulation, and one-dimensional simulation based on correlations defined in ISO 15099 through the WINDOW software developed by LBNL.

The LBNL developed simulation program, WINDOW, calculates heat transfer through glazing assemblies based on one-dimensional correlations of average values. The correlations were developed through a variety of means including laboratory measurements, detailed simulations, and analytical models. The correlations and their sources are defined in ISO 15099. Of specific interest for this work is the model presented for thermally driven ventilation. The model is based on introducing a modifier to the unvented correlation to determine a modified average surface-to-air heat transfer coefficient for the between glass space. The result is solved for iteratively as presented in Figure 1. The modifying equations are based on several pressure drop principles for air through a channel including Bernouilli, Hagen-Poiseuille, and entrance/exit correlations. While these equations are physically based, the standard does not cite any validation of the approach through either measurements or detailed simulation.

Steady State Measurements

Industry standard quantitative measurement of window heat flow is performed with calorimetric "hot box" instruments as outlined in ISO 12567 (ISO 2010). This method is effective for overall thermal performance but gives little localized information. For this reason, the Windows and Envelope Materials Group at Lawrence Berkeley National Lab (LBNL) has developed a technique for quantitative thermal measurement of windows using IR thermography. The IR thermography technique described by Griffith et al. (1998) and Griffith et al. (2001) allows us to measure window thermal performance with much more accuracy than the ISO test while obtaining detailed information used to verify LBNL's simulation algorithms. The chamber is typically configured to closely match NFRC 100 (2010) boundary conditions of Tc = -18 °C, hc = 26 W m-2 K and Tw = 21 °C, hw determined by natural convection, where Tc and Tw are the temperatures on the cold and warm sides of the test chamber and hc and hw are the surface-to-air heat transfer coefficients. Figure 2 illustrates the test chamber configuration.

Experiments were performed using specially instrumented specimens representing the typical thermal resistances of a glazing system and a shade layer. Both the glazing layer and the shading layer are constructed like calibrated transfer standards (CTS), samples of known thermal resistance with many temperature pair measurements that allow calculating the local heat flux through the specimen in many locations. Thermocouples are placed in a 3 x 6 grid (18 per surface) on the inside surface of two 1080 mm by 775 mm glass panes with a calibrated foam layer between. Figure 3 illustrates the configuration of the CTS specimens utilized for this study and the locations of the sensing thermocouples, TCTS. The measured temperatures are taken to represent the average temperature of the immediate area, A, surrounding each sensor.

Principal Investigator(s)