Advanced Computer Modeling
UL is using innovative techniques to understand and predict how materials behave. We focus on the behavior and structural response of steel fire doors and wood beams in a fire environment.
WHY ADVANCED COMPUTER MODELING MATTERS
By using advanced computer modeling, UL is able to effectively leverage physical and virtual testing. Physical testing in the study of fire is expensive, and thus, generally very few tests are run on a particular product. Also, physical testing can only provide data at discrete points. However, it is still very important to design and carry out physical tests, as they can be valuable in validating computer models. Once a computer model is validated, it can be run under a variety of different input parameters much more economically than physical testing.
Under pressure from fast and furious product development cycles and more stringent fire standards, manufacturers that wait for physical testing to find out if their products will perform safely may suffer a competitive disadvantage. Add to this mix that several new technologies are being developed such as fire-resistant coatings and engineered composites, and the challenge is enormous. UL now has the expertise in applying advanced computer modeling to complement fire testing and help establish an efficient hybrid approach to advancing Fire Safety science, fire standards and product safety.
UL has been using these advanced engineering tools to create virtual models of some products in fire tests within the standards. With such models, insights into the behavior of products in the extreme environment of fire can be gained, and when validated, a model can be adjusted to more efficiently predict the outcome of multiple scenarios and product design variations.
UL has been using computer modeling related to the Fire Safety of products in several ways:
- The performance of building components and materials subjected to a fire environment using a coupled thermal-structural finite element solver (ANSYS).
- Computational fluid dynamics (CFD)-based fire modeling software to predict the performance of fire sprinklers including interaction with fire.1
- Modeling the thermal performance of spray-applied fire-resistant coatings on steel columns for a variety of column shapes and sizes and coating thicknesses.
In all these cases, modeling provides a data-rich output, which can be analyzed and visualized in multiple ways to help provide the necessary insight to understand Fire Safety risks.
WHAT DID UL DO?
To provide a better picture of how UL is using state-of-the-art computer modeling, two important and challenging examples of the fire performance of building components will be covered: (1) the behavior of steel fire doors subjected to the fire endurance test and (2) the behavior of engineered wood products in a fire environment.
Both examples were analyzed using the finite element analysis (FEA) methodology. To build and solve an FEA model requires specific information describing the material properties, boundary conditions, assembly geometry and constructional details, loadings and even some consideration of the possible failure mode(s).
PREDICTING THE BEHAVIOR OF STEEL FIRE DOORS
Fire doors within a building are meant to resist the spread of fire from one part of a structure to another, to enable more effective fire mitigation and safe egress of the occupants. As a means of evaluating fire resistance, fire door assemblies are tested according to standards such as UL 10B.2 To effectively and efficiently build a suitable finite element representation of a fire door assembly subjected to a fire exposure, it is prudent to assess the necessary amount of detail that should be captured. UL has studied fire door modeling over the years 3 and has found a variety of design details that are critical to accurate predictions.
However, first UL identified key engineering assumptions, based on our long history of testing fire doors, which generally hold true and help guide the finite element analysis. These include:
- The wall and frame holding the fire door are rigid during the entirety of the test and so need not be modeled in detail.
- The thermal insulation does not provide any structural stiffness to the fire door assembly.
- The coupling between the thermal and structural responses is one-way during the early parts of the test as the structural response has a negligible effect on the thermal response.
Without describing the extensive fire door modeling work that UL has carried out in this area, we focus on one key aspect. For proper modeling of fire doors, the thermal contact resistance between mating steel components is critical. 4 Generally steel fire doors consist of steel panels and steel stiffeners. These parts are structurally connected through welds. As the door deforms, these parts could deform differentially, changing the thermal resistance. Our research found that simply assuming that heat transfer occurs through the weld points would underestimate the heat transfer through the door and in turn affect the predicted structural results. Information about the thermal contact resistance and its changing nature is generally not known by the manufacturer. However, the model can show the effect of different thermal contact configurations and provide insight into design decisions that may improve the fire performance of the door.
THE FIRE PERFORMANCE OF WOOD BEAMS
This modeling work was part of a larger multi-year research plan to understand the fire performance of engineered wood, common in new residential dwellings versus traditional lumber that is typical of older homes. Despite the large amount of testing, a validated FEA model of wood beams that can predict the effect of different design changes would be very valuable in developing a strong technical basis for possible building code revisions and changes in firefighting tactics.
Unlike other building components such as masonry and brick, there are several specific challenges in predicting the response of wood-based structures to fires. The key challenge is that wood burns. The burning of wood leads to material degradation and decomposition through pyrolysis. Wood is also a complex composite of natural polymers and is generally anisotropic, heterogeneous and porous. The properties of wood are also affected by moisture content. And in a fire environment, any moisture contained within wood will evaporate and diffuse, altering material properties. Last, the failure mode of a wood-based building component would depend upon details of the construction, material imperfections, connections, etc.5 What’s more, all these material properties would need to be known over the temperature range that wood would reach in a fire environment.
Our modeling results were able to demonstrate quantitative agreement in trends seen during testing for both single beam and wood beam floor assembly tests. The key trend was that the traditional lumber beam lasted much longer than the engineered wood beam under similar fire and mechanical loading conditions. In addition, the model provides insight into parameters such as charring rate, which was found to compare favorably with the range of data in the published literature.6
The FEA model and advanced analysis were able to predict the onset of instability where the deflection rate increased substantially. The model also revealed that for the engineered wood beams, the main failure path is the burnout of the web, thereby transferring load sharing to the top chord, as the lower chord — though mostly unburned — was then separated. For the traditional lumber rectangular section beam, the failure path mainly reduces cross section through three-sided heating and through a combination of weakened material properties and reduced cross section, which eventually fails to sustain the load.
UL continues to innovate using our expertise to offer a more comprehensive approach to Fire Safety. By using computer modeling and sophisticated analytical techniques, we are more effectively able to predict fire behavior, assess risk and provide insights that can lead to better design decisions.