Environment
PNNL environmental research aims to transform environmental science from reactive to predictive, influencing environmental decisions by enabling the prediction of the effects of human activities on natural systems. The Environment application area includes modeling and simulation in climate, fluid dynamics, hydrology, groundwater transport, and remediation products and processes.
Atmospheric and Climate Modeling
The climate is known to change in response to changes in the planetary energy balance, and emissions of carbon dioxide and sulfur dioxide from combustion of fossil fuels are perturbing Earth's energy balance. Empirical methods of estimating the climatic impact of these emissions are limited because the radiative forcing is difficult to measure and is unprecedented in the historical record. Thus physically based climate models, after they are shown to be reliable, are used as powerful and flexible tools for predicting the likely climatic response to past and future emissions of carbon dioxide and sulfur dioxide.
Global climate models are based on known conservation laws for mass, momentum, energy, and water. Because they require time steps of minutes but are used to predict climate change on time scales of months to centuries, their horizontal spatial resolution is typically at least 100 km; hence their treatment of physical processes is approximate.
PNNL's global climate model is based on the National Center for Atmospheric Research (NCAR) Community Climate Model (CCM2). In addition to the CCM2 physics, the model features the following:
- bulk cloud microphysics (cloud water, cloud ice, rain, and snow)
- droplet number prescribed or predicted from aerosol activation
- cloud radiative properties related to particle size
- aerosol radiative properties related to composition, size, and relative humidity
- prognostic turbulence kinetic energy
- optional nudging toward observed winds.
Global models cannot currently afford the spatial resolution needed for climate impact assessment. To provide adequate resolution within a limited domain, regional climate models have been developed. These models can be driven either by observed meteorology on their lateral boundaries or by meteorology that is simulated by global climate models.
Computational Fluid Dynamics
Computational fluid dynamics (CFD) simulation is applied to a variety of projects at PNNL for multiphase flows, fuel cell research, and special flow geometries.
Direct simulation of multiphase flow remains a challenging problem for CFD. The need to explicitly model the dynamics of the interface between different phases and the associated problems of adjusting the computational grid used to compute the flow within each phase present tremendous challenges for conventional CFD approaches. Over the last decade, the Lattice-Boltzmann method has been developed as an alternative to conventional CFD, showing great promise for simulating multiphase flows and flows in topologically complex media. This approach is based on a discretization of the Boltzmann equation. In contrast, most CFD methods are based on a discretization of the continuum hydrodynamic field equations.
Fuel cells will play an important role in future energy production, and an integrated set of modeling tools is invaluable to their design. Researchers at PNNL are developing computer models to provide a unified method for analyzing the fluid, thermal, electrochemical, and structural response of the cells. Computational fluid dynamics analyses solve the flow and thermal problems associated with this research. Close coupling with the finite element analysis determines the corresponding stress levels and safety factors. In addition, specialized software being developed by PNNL researchers will predict fuel cell performance and assess the reliability and lifetime of the cell.
Models for CFD readily adapt to new geometries and typically include a collection of physics models. The Laboratory's TEMPEST code is currently being used for modeling chemically reactive flow in radioactive waste tanks and Joule-heated glass flow in waste glass melters. The commercial CFD code STARCD is being modified and applied to the complex flow in cold-crucible melters for reactive metals and in dynamic behavior in solid oxide fuel cell stack operation.
Groundwater Transport Modeling
Computer models are used to forecast future groundwater conditions and predict the movement of contaminants in groundwater. Such predictions are important, for example, in planning waste management and cleanup activities for the Department of Energy's Hanford Site. A site-wide numerical model of groundwater flow and contaminant transport has been developed and is being improved and refined. Local-scale modeling is used to predict the migration of dense, nonaqueous liquid disposed of in waste cribs through the vadose zone and aquifer. Local-scale modeling has also been used for the past several years to design and evaluate pump-and-treat systems for local-scale groundwater contaminant plumes.
Hydrology Modeling
The hydrology research activities at PNNL cover a broad range of water resource issues, including these:
- assessing the impacts of climate change, hydropower, and irrigation on watershed health
- predicting the movement of fluids, gases, and contaminants through the soil and vadose zone above the water table
- analyzing complex problems involving the impacts of seasonal water availability, temperature, suspended sediment, and dissolved gases in rivers systems, especially the impacts of dams and their operation.
One of the more serious potential consequences of greenhouse warming is a change in the phase of wintertime precipitation in temperate regions. Even if the amount of precipitation does not change, the expected higher proportion of precipitation falling as rain rather than snow in a warmer climate has serious implications for management of water resources. In the western United States, where about 70% of the annual runoff is derived from snowmelt in headwater catchments, a higher fraction of wintertime precipitation falling as rain implies reduced storage of water in the winter snowpack and hence a higher likelihood of both wintertime flooding and diminished water resources during summer unless additional artificial reservoirs are constructed. Because these impacts could lie well beyond the range of those in recorded history, physically based models are needed to estimate the impact of greenhouse warming on water resources in mountainous regions.
Product and Process Modeling
Solving waste remediation and other environmental problems requires models and simulations for properties of treatment/remediation products (such as nuclear waste glass) and processes. PNNL has been involved in this research for decades to solve environmental problems at Hanford and numerous other sites.
Treatability studies are conducted to help characterize the physical, chemical, and microbiological nature of liquid and solid waste streams and devise effective, economical ways to treat and manage such wastes to meet the regulatory criteria for safe disposal or reuse. These studies are necessary to determine specific treatment and recycling technologies as well as capital and operating costs. Using experiments and computer models in tandem, these studies characterize the quality of waste streams and the design processes to treat them effectively.
Statistical experimental design and modeling methods have been developed and applied over the last 25 years at PNNL to predict properties of nuclear waste forms (such as glass and grout) as functions of composition and temperature. In particular, special methods for the experimental design and modeling of mixture experiments (in which the proportions of components in the end product must sum to one) have been developed and applied to waste form development problems. These methods yield empirical and semi-empirical models that tend to predict better than theoretical models because of the large number of waste components and additives that affect waste form properties such as chemical durability, viscosity, and electrical conductivity. The models are used to optimize waste form compositions and can be used to control waste immobilization processes and demonstrate compliance with applicable requirements. Also applied are statistical methods to quantify uncertainties of waste-form property models and Monte Carlo simulation methods to propagate processing uncertainties through the models and mass-balance equations.
High performance computing systems are being used in chemical process modeling to minimize the construction and operational costs in facilities built to treat the Department of Energy's legacy of stored wastes and to maximize worker safety.
The inherent uncertainty in site characterization and physical process modeling makes computer simulation indispensable in evaluating and tuning various groundwater contamination and remediation scenarios and risk factors. Such applications, however, present great challenges for meaningful simulation. For example, performance assessment of nuclear waste repositories requires simulation of flow, transport, dispersion, and nuclear and chemical reactions over extraordinarily long time periods to extraordinary accuracy—parts per trillion for some highly toxic species.
Risk Assessment Modeling
Environmental risk and impact assessments are necessary during various stages of project implementation to ensure that the environmental implications of decisions are taken into account before the decisions are made. The process, required by several regulatory agencies and governments, involves analyzing the likely effects on the environment, recording those effects in a report, eliciting public comment on the report findings, taking into account the comments and findings when making the final decision, and informing the public about that decision afterwards. For many years, PNNL scientists have conducted human and ecological risk assessments for federal agencies. PNNL has developed a methodology to evaluate ecological risks from various remedial alternatives and ensure that the remediation procedure is not more destructive to the environment than existing contamination.
The Nuclear Regulatory Commission Spent Fuel Project Office has maintained a close relationship with PNNL modelers running computational fluid dynamics, thermal analysis, and structural dynamics codes as they seek to better understand the risks to the environment from engineered systems. Model developers modify and apply a suite of codes such as COBRA-SFS, TEMPEST, ANSYS, and LSDYNA, typically exceeding past applications in rigor.