Dr. Gregory S. Jackson
Associate Professor
Ph.D. Cornell University 1994

Hazard Assessment of Large-Scale Releases of Combustible Chemicals (Sponsored by ONR)

Click here to view a PowerPoint presentation

Team members:
Greg Jackson (University of Maryland, College Park); Arnaud Trouvé (University of Maryland, College Park); Tom McGrath (NSWC – Indian Head Division); Bill Hinckley (NSWC – Indian Head Division)

Background

Transportation and storage of large amounts of toxic and/or reactive chemicals present security and safety risks associated with accidents or malicious attacks. For reactive chemicals, uncontrolled energy release in a large-scale fire or detonation can be catastrophic as demonstrated by the September 11th tragedy and the 1947 Texas City disaster. The enormous risks associated with reactive chemicals have led to heightened concern over issues such as transport of anhydrous ammonia and installation of large LNG facilities. Numerous simulation tools have been developed by government agencies and commercail interests to assess associated risks of large-scale release scenarios of hazardous chemicals. These tools include rleatively simple computationally fast models appropriate for rapid hazard assessment by first responders and field managers as well as more advanced models comprehensive scenario simulation by analysts. These existing codes incorporate Gaussian turbulent dispersion models with semi-empirical or semi-analytical models for source terms and subsequent reaction-based thermal energy releases.

Motivation

While existing first-responder models provide valuable approximations for toxic exposure and/or damages due to thermal energy release or detonation overpressures, they are based on open-field calculations and simplistic, often semi-empirical, models for source terms and combustion rates. They can substantially under- or over-predict the extent of exposure or damage, particularly in congested urban environments. In first response applications, such errors can result in either tragic loss of life or excessive disruption and subsequent public distrust of emergency management teams. With continued advancement in computational fluid dynamics for turbulent flows and in experimental diagnostics for complex and reacting flows, the aim of this effort is to improve existing risk assessment tools by providing higher fidelity sub-models and by incorporating advances in turbulent flow modeling to understand the iimpact of structures and congestion on chemical dispersion and possible subsequent energy release.

Goals

This project is a collaborative effort between the CECD and the Naval Surface Warfare Center, Indian Head Division (NSWC IHD). The goals of the project are to advance first responder and safety/security analyst risk assessment of large-scale chemical and biological releases. The strategies to purse are the following:

• experimentally evaluate and refine source models for large-scale liquid spills and multi-phase releases from cryogenic and other liquid storage systems
• develop and experimentally test improved energy release models for large scale fire scenarios, detonations of reactive fuels, and for boiling liquid explosions (BLEVE's)
• advance sub-models for interactions between large-scale turbulent flows and structures, and build a database from subsequent simulations of dispersion scenarios in congested environments
• develop improved models for structural response to thermal energy absorption and detonation overpressures and compare with historical records for model evaluation
• work with interested code developers to incorporate model advances into existing first responder codes for risk assessment of large-scale chemical releases
• provide advanced detailed simulations of selected release scenarios in congested environments for training emergency response teams.

Results

Today, the model development effort has yielded a preliminary working version of the code suite and proof of concept simulations in large (order of 100 meters) 3D computational domains. The model development efforts have yielded significant computational advancements over traditional modeling approaches, and increased iinsight into phenomena critical to predicting large-scale chemical dispersion and explosion/fire events. Experimental support for the modeling effort, including large fuel/air detonability tests and lab-scale spill and dispersion experiments, have provided necessary data for validating the computational models, as well as valuable insight into dispersion and detonation phenomena. While the work to date has yielded a preliminary working version of the code suite, more work is needed to ready the model for real-world use.