John W. Labadie

Neural-optimal control algorithm for real-time regulation of in-line storage in combined sewer systems

Attempts at implementing real-time control systems as a cost-effective means of minimizing the pollution impacts of untreated combined sewer overflows have largely been unsustained due to the complexity of the real-time control problem. Optimal real-time regulation of flows and in-line storage in combined sewer systems is challenging due to the need for complex optimization models integrated with urban stormwater runoff prediction and fully dynamic routing of sewer flows within 5-15min computational time increments. A neural-optimal control algorithm is presented that fully incorporates the complexities of dynamic, unsteady hydraulic modeling of combined sewer system flows and optimal coordinated, system-wide regulation of in-line storage. The neural-optimal control module is based on a recurrent Jordan neural network architecture that is trained using optimal policies produced by a dynamic optimal control module. The neural-optimal control algorithm is demonstrated in a simulated real-time control experiment for the King County combined sewer system, Seattle, Washington, USA. The algorithm exhibits an effective adaptive learning capability that results in near-optimal performance of the control system while satisfying the time constraints of real-time implementation.

Enhancing water quality in hydropower system operations

The quality of impounded waters often degrades over time because of thermal stratification, sediment oxygen demands, and accumulation of pollutants. Consequently, reservoir releases impact water quality in tailwaters, channels, and other downstream water bodies. Low dissolved oxygen (DO) concentrations in the Cumberland River below Old Hickory dam result from stratification of upstream reservoirs and seasonally low release rates. Operational changes in upstream hydropower reservoirs may be one method to increase DO levels without substantially impacting existing project purposes. A water quality model of the upper Cumberland basin is integrated into an optimal control algorithm to evaluate water quality improvement opportunities through operational modifications. The integrated water quantity/quality model maximizes hydropower revenues, subject to various flow and headwater operational restrictions for satisfying multiple project purposes, as well as maintenance of water quality targets. Optimal daily reservoir release policies are determined for the summer drawdown period which increase DO concentrations under stratification conditions with minimal impact on hydropower production and other project purposes. Appendixes A–D available with entire article on microfiche. Order by mail from AGU, 2000 Florida Ave., N.W., Washington, DC 20009 or by phone at 800-966-2481;

Multicriteria Decision Support System for Regionalization of Integrated Water Resources Management

2.50. Document W97-003. Payment must accompany order.

Optimal Allocation of Short-Term Irrigation Supply

Successful implementation of integrated water resources planning and management (IWRM) requires delineation of regions that are relatively homogeneous with respect to multiple criteria, including hydrographic, physical-environmental, socioeconomic, and political-administrative aspects. The water resources planning and management (WARPLAM) DSS is presented as tool for regionalization in support of IWRM through: (1) GIS processing of spatial data related to multiple criteria for defining the homogeneity of clustered base units (e.g., catchments) with respect to multiple criteria; (2) application of fuzzy set theory to development of composite measures of homogeneity over all criteria for alternative clustering of adjacent base units; and (3) development of a modified dynamic programming clustering algorithm that guarantees consistent optimal solutions based on user preferences on the relative importance of the suite of criteria considered for regionalization. The viability of WARPLAM DSS as a tool for regional delineation in support of IWRM is demonstrated through a case study application to the Tocantins-Araguaia River Basin, the second largest in Brazil.

Generic Simulation Models for Facilitating Stakeholder Involvement in Water Resources Planning and Management: A Comparison, Evaluation, and Identification of Future Needs

An optimization model has been developed toaid decision making in real time fordeficit irrigation when conflict betweenwater supply and demand arises in amultiple crop irrigation scheme. The resultis the optimal allocation of short-termsupply of irrigation water. Theoptimization model is based on DynamicProgramming. In the optimization model, theshort-term supply is optimized in functionof a specified strategy determined by theuser. The strategies that the user canselect from are: maximum benefit, equitablebenefit, equitable yield and maintainingequity in the system. The potential of themodel has been assessed through applicationof the model to Perkerra irrigation schemein Kenya. In the 680 ha scheme, maize,onion and chili are cultivated in theirrigation season. Analysis of the resultsfor the 1999/2000 season, where the watersupply was 35 percent smaller than thedemand, indicates that improvements in cropproduction can be achieved throughapplication of the optimization model.Sensitivity of production system to variouslevels of water restriction is demonstratedby sensitivity analysis.

Optimal On-farm Allocation of Irrigation Water

Abstract The keen public interest in the management of water resources and related environmental and socioeconomical issues has placed a great demand for a new class of generic simulation models capable of facilitating the involvement of stakeholders at all levels in the decision-making process. This chapter presents the results from a workshop on the assessment of the current state of the science and art in river basin simulation modelling. It also identifies future needs from the perspective of those involved in the development and/or the application of five leading generic river basin simulation models. The models discussed in this chapter have many features in common, and some that differ. All are what we commonly call decision support systems for analysing and estimating various impacts associated with alternative water, and, in some cases, watershed management policies and practices. All are interactive, i.e. they use graphic interfaces and are menu driven. Each model is designed to be flexible enough to apply to most if not all river basin systems no matter where located. Each is a tool that can be used, at least as a first step, in the identification and evaluation of preferred management policies and practices. Just how well models like these have met their goals in their many applications, and just how well they have facilitated stakeholder involvement? This question was the subject of one of the International Environmental Modelling and Software Society (iEMSs) workshops in 2006. The purpose of the workshop was to identify weaknesses and research and development needs that can lead us toward more effective models for addressing current and potential future management issues. This chapter is based on the results of the workshop. The five models that were discussed in this workshop included MIKE BASIN from Danish Hydraulics Institute (DHI) in Denmark, MODSIM from Colorado State University in western US, RIBASIM from Delft Hydraulics in The Netherlands, WBalMo from WASY (now part of DHI) in Germany, and WEAP from Stockholm Environmental Institute in Sweden and the US. Several participants in the workshop had experiences developing and using other generic river basin simulation software.

Decision Support System for Adaptive River Basin Management: Application to the Geum River Basin, Korea

ABSTRACT AN irrigator must make many decisions in determining the date and volume of water to apply to each field. Scheduling programs that rec-ommend the date and volume have been developed and used for assisting management. The new management program optimizes the on-farm alloca-tion of water within the limits of avail-able water and labor as demonstrated on one farm with 40 fields.

Optimal integrated operation of reservoir-assisted stormwater treatment areas for estuarine habitat restoration

Abstract The MODSIM 8.0 decision support system (DSS) for integrated river basin management (IRBM) has been adapted from a prior appropriation rights-based system to one found in Korea and in much of Asia where water deficits are shared among water use sectors, taking into account priorities established by water policy and institutional frameworks. The Korean version called KModSim is applied to the Geum River basin for evaluation of long-term sustainability of existing and new water infrastructure and facilities under integrated, basin-wide water resources management. KModSim is calibrated to the physical and hydrologic characteristics of the basin, as well as to operational and administrative water allocation policies for municipal and industrial water supply, irrigation, hydropower, transbasin diversions, and low-flow augmentation for environmental purposes. Conditional reservoir operational rules that adapt to changing river basin hydrologic conditions are developed from an implicit stochastic optimization algorithm and incorporated using the extensive user-customization capabilities of KModSim. Results demonstrate that decision guidance under KModSim enhances beneficial water uses in the Geum River system through fully integrated, basin-wide management.

Combining Simulation and Optimization in River Basin Management

Significant reductions in pollutant loadings can be achieved through coordinated operation of constructed wetlands or stormwater treatment areas (STAs) when connected to retention reservoirs for maintaining ideal water levels and hydraulic residence times in STAs that maximize treatment efficiency. Reservoir-assisted STA systems can provide additional benefits including maintenance of target frequency distributions of stormwater inflows into coastal estuarine systems for ecosystem rehabilitation, dry season flow augmentation in coastal riverine systems for maintaining the lower salinity zone, and supplemental water supply. An implicit stochastic optimization approach is applied to adaptive, multiobjective control of interconnected reservoir-assisted STAs that optimizes fuzzy rule-based operating rules using a genetic algorithm interacting with a simulation model of the stormwater drainage system. The model is applied to the four interconnected reservoir/STAs in the design configuration of the Indian River Lagoon-South project for the St. Lucie River watershed and estuary, South Florida. Results of simulated long-term performance of the optimal operating rules show attainment of targets for remediation of the aquatic ecology of the St. Lucie estuary, riverine environmental protection, and reliable supplemental water supply.

Optimal Short Term irrigation schedules

Historically, there has existed a dichotomy in relation to the use of descriptive models (i.e. simulation) and prescriptive models (i.e. optimization) in river basin management. This is unfortunate, since there are unique advantages with each modelling approach which can be greatly enhanced through their joint usage. Descriptive models allow accurate modelling of complex flow and storage mechanisms in a river basin, and are conducive to Monte Carlo analysis and assessment of risks. Descriptive models, on the other hand, cannot directly find the “best” solutions, which represents the primary advantage of prescriptive models. The latter, however, often require simplifying assumptions on model structure. A combined simulation-optimization strategy for river system management is presented. The procedure is evaluated and tested using the Nizao River system in the Dominican Republic as a case study.