By Larry Deschaine, Ph.D., P.E., Tad Fox, PG, and Jeffrey Fairbanks
Water planning and management efforts often involve evaluating complex, interconnected hydrologic systems. Assessing potential locations for new groundwater supplies might require examining the long-term effects of increased pumping of existing domestic, municipal, agricultural, or commercial wells. In some cases, groundwater and surface water or wetlands may be strongly interconnected. The impacts on habitat and water quality must be considered. Computer simulation models have been invaluable in assessing specific water supply and water management alternatives.
Decision-makers must balance potential hydrologic effects with cost, management, scheduling, and institutional factors to identify the solution that will best meet project objectives and address stakeholder concerns. Efforts to optimize the assessment of alternatives typically employ qualitative methods, such as using a computer model. However, the variety of decision variables (for example, the number of new wells, the range of possible pumping rates at each well, or the number of possible new well locations) and identified constraints leads to millions of potential combinations. Consequently, the ability to determine an optimal solution using such techniques is limited by the cost and time to perform the modeling and assess the results. This significantly reduces the number of scenarios that can be evaluated.
PBMO has been used for supporting optimal water resource management for municipalities, for dewatering excavation and mining sites, and for managing and remediating contaminated water resources at government and commercial facilities.
New parallel optimization technology deployed in cloud-based, highperformance, computing environments is providing more efficient, effective, and comprehensive decision support tools for water planning and management. One such innovation is Physics-Based Management Optimization (PBMO). This technology finds the best solution within the physical constraints of the problem to a userspecified set of objectives, such as cost, schedule, and pumping locations.
PBMO has been used for supporting water resource management for municipalities, dewatering excavation and mining sites, and management and remediation of contaminated water resources at government and commercial facilities. The technology integrates optimization techniques with data models, expert systems, and traditional engineering savvy with high-performance distributed computing (internal networks or cloud resources).
TECHNOLOGY AT WORK
PBMO can be conceptualized as a medallion with two interdependent and complementary halves. One half contains a module with calibrated physics-based models that represent the physical system. This half seamlessly interacts with the second half, which houses the computational optimization module. The software architecture includes a parallel generalized global optimizer coupled with a candidate solution search engine capable of executing 5,000 model runs in parallel batch mode. The search strategy is repeated until the optimal solution is determined.
Decision-makers specify the desired project objectives and constraints as inputs. PBMO then performs the computational optimization by using a suite of embedded techniques to iteratively and in parallel communicate between the model module and the optimization module to find the solution that best satisfies those objectives while honoring the constraints.
For instance, possible objectives may include reaching a specific total pumping rate at the lowest cost and ensuring sustainable operation with constraints on the number of new wells and where they can be installed. By directly linking numerical modeling and computational optimization, PBMO concurrently assesses the impacts of site-specific, physics-based hydrologic interactions on planning and management factors such as cost, schedule, and land use restrictions. This approach eliminates the subjective judgments incorporated into traditional optimization methods and results in structured, credible solutions acceptable to stakeholders and regulators.
By taking advantage of distributed computing on internal networks or in the cloud, PBMO can rapidly evaluate tens of thousands of potential engineering options and provide transparent comparisons as output. Because all potential solutions have been evaluated computationally and ranked according to quantitative priorities, project managers and stakeholders can have high confidence in selecting alternatives.
The technology permits evaluation of many hundred, to many thousands, of potential solutions in the same amount of time that conventional methods require to evaluate only a limited number.
PBMO also identifies cases where the existing constraints on the desired planning or management objective are such that a viable solution is not possible. This is valuable information, especially if the problem formulation was driven in part by regulatory or other institutional requirements that could not be met.
Groundwater Restoration/Containment. PBMO has been utilized on Department of Defense and commercial sites to optimize existing contaminated groundwater remediation systems and to design new systems to minimize the volume of water pumped/treated and to contain/ prevent plume migration to drinking water resources. The containment aspect is particularly applicable for preventing groundwater impacted by emerging contaminants from migrating beyond the boundaries of a military installation. For cost and sustainability calculations such as reducing carbon footprint, or minimizing impacted resources, the technology includes an optimization user interface that allows tracking of solution performance and quantification of cost and energy components associated with each viable solution.
Municipal Water Supply. Planning efforts typically involve ensuring that future water needs will be met sustainably while minimizing adverse effects on adjacent users. For a groundwater management area in Texas, two scenarios were evaluated using PBMO as part of a regional water planning effort to determine the maximum possible pumping rate for a multi-layer and multi-basin aquifer system. The first scenario evaluated the maximum pumping rate that could be achieved county-wide using more than 1,000 wells while meeting specified constraints on the maximum total extraction rate per year and the maximum permissible drawdown by aquifer at any location. The second scenario was similar, but evaluated the maximum pumping rate attainable by basin under similar management constraints.
As part of another municipality’s strategic plan, PBMO was used to support decisions on how best to control a nitrate plume that had resulted from the beneficial reuse of treated wastewater. In this case, the treated wastewater was used for irrigation purposes. The city’s strategic plan included continued beneficial reuse of the treated effluent. PBMO was deployed to identify optimal plume containment solutions in order to reduce off-site migration of impacted groundwater and support longer-term decisions for optimal groundwater remedy design.
Dewatering Strategy Design. Excavation dewatering is another potential component of water planning and management, particularly when extracting mineral resources from below the water table. In this case, PBMO was used to design a dewatering strategy for two adjacent iron ore lenses at an open pit mine in Australia. The problem was complex, with many possible dewatering options and schedule constraints. The model used was a nonlinear variably saturated flow model previously developed by the mine operator to prepare a dewatering plan for a 12-year period. PBMO yielded a solution that reduced the number of wells and total pumping volume as identified in the plan. The PBMO application resulted in an estimated savings of $4.5 million in well installation costs, reduced the volume of water pumped by 4.1-billion-gal, reduced water extraction costs by $4.6-million, and met the project schedule milestones.
Flexible Applications. The modular nature of the technology allows surface water modeling tools to be integrated into the lower half of the medallion and facilitates optimal design of engineered solutions for surface water planning and management efforts like sustainable stream withdrawals or climate change considerations. Additionally, variable density tools have been integrated into PBMO to support water planning and management efforts in coastal areas where sea water intrusion is an issue.
Larry Deschaine, Ph.D., P.E., is Principal Engineer, Tad Fox, PG, is Director of Applied Modeling, and Jeffrey Fairbanks is Director of Software Sales, HydroGeoLogic Inc. They can be reached at firstname.lastname@example.org; email@example.com; and firstname.lastname@example.org
[This article first published in the July-August 2018 issue of The Military Engineer]