NUMERICAL SIMULATION OF GROUNDWATER FLOW AND SOLUTE TRANSPORT IS AN ESSENTIAL ELEMENT OF ANY ASSESSMENT OF THE PERFORMANCE OF A PROPOSED REPOSITORY FOR NUCLEAR WASTE. THE ACCURACY AND REALISM OF SUCH SIMULATION MUST BE CLEARLY DEMONSTRABLE AND ALSO DEFENSIBLE AGAINST CRITICAL PUBLIC SCRUTINY. DETAILED SIMULATIONS ARE EXPENSIVE SO ANY INCREASED COMPUTATIONAL EFFICIENCY OFFERS THE OPPORTUNITY FOR MORE EXTENSIVE SENSITIVITY ANALYSIS AS PART OF A COMPREHENSIVE REPOSITORY ASSESSMENT.
THE PROJECT AIMS TO IMPROVE THE CAPABILITY, EFFICIENCY AND REALISM OF THE NAMMU CODE, WHICH SIMULATES GROUNDWATER FLOW AND SOLUTE TRANSPORT THROUGH A POROUS MEDIUM. OUR DETAILED OBJECTIVES ARE TO DISCOVER AND EXPLOIT SUPERIOR METHODS FOR SOLVING SIGNIFICANTLY NON-LINEAR PROBLEMS; TO INTRODUCE INTO THE CODE A CAPABILITY FOR MODELLING CHEMICAL REACTIONS; TO IDENTIFY AND TEST IMPROVED TECHNIQUES FOR SIMULATING THE PROGRESS OF SHARP FRONTS IN SOLUTE CONCENTRATION; AND TO IDENTIFY BETTER-FOUNDED REPRESENTATIONS OF SOLUTE DISPERSION. THE LAST TWO TOPICS WILL BE COVERED BY SUB-CONTRACTS PLACED AT UNIVERSITIES IN THE UNITED KINGDOM.
THE NAPSAC CODE PRESENTLY CALCULATES FLOW THROUGH A THREE-DIMENSIONAL NETWORK OF FRACTURES. OUR AIM IS TO DECIDE UPON AND IMPLEMENT THE MOST APPROPRIATE METHOD FOR CALCULATING SOLUTE TRANSPORT THROUGH A THREE-DIMENSIONAL NETWORK, SO NAPSAC COULD BE USED TO INVESTIGATE RADIONUCLIDE MIGRATION FROM REPOSITORY IN HARD FRACTURED ROCK.
The aim of this project was to improve the capability, efficiency and realism of the NAMMU and NAPSAC codes, which simulate groundwater flow and solute transport.
Using NAMMU, various solution methods for nonlinear problems were investigated. The Broyden method gave a useful reduction in computing time and appeared robust. The relative saving obtained with this method increased with the problem size. This was also the case when parameter stepping was used.
The existing empirical sorption models in MANNU were generalized and a ternary heterogeneous ion exchange model was added. These modifications were tested and gave excellent results. The desirability of coupling NAMMU to an existing geochemical speciation code was assessed.
Numerical methods for the solution of cases involving sharp fronts in solute concentration were investigated. Roe's Superbee method was applied to several test cases. The method gave impressive results for some high Peclet number test problems. However, problems were encountered in cases with large permeability contrasts.
Models of solute dispersion were investigated. A survey of the literature, and a Monte Carlo study of the early stages of plume development were carried out. The concentration distribution did not settle down to a normal distribution within the length and timescales of the experiment. Methods of calculating effective parameters were investigated.
A very efficient particle following algorithm has been added to the NAPSAC code, enabling tracer transport to be predicted through large fracture networks. The new algorithm has been tested against 3 test examples. These demonstrations confirmed the accuracy of the code for simple networks, where there is an analytical solution to the transport problem, and illustrated the use of the computer code on a more realistic problem.
1. IMPROVED METHODS FOR SOLVING SIGNIFICANTLY NON-LINEAR GROUNDWATERFLOW PROBLEMS WILL BE IMPLEMENTED IN NAMMU BECAUSE CURRENTLY AVAILABLE, ALTHOUGH ROBUST, ARE EXPENSIVE.
2. THE NAMMU CODE WILL BE ENHANCED TO PROVIDE A MEANS OF MODELLING THE IMPACT OF CHEMICAL REACTIONS ON SOLUTE TRANSPORT IN A POROUS MEDIUM.
3. METHODS WILL BE ASSESSED AND TESTED THAT OFFER PROSPECTS OF SIMULATING EFFICIENTLY THE ADVANCE OF SHARP FRONT IN SOLUTE CONCENTRATION, WITHOUT SPURIOUS DISPERSION.
4. MODELS OF INHOMOGENEOUS MATERIALS WILL BE EXPLORED TO PROVIDE BETTER UNDERSTANDING AND DESCRIPTIONS OF SOLUTE DISPERSION.
5. THE NAPSAC FRACTURE-NETWORK CODE WILL BE ENHANCED BY INCORPORATING A JUDICIOUSLY CHOSEN TECHNIQUE FOR CALCULATING SOLUTE TRANSPORT.