Search

In wastewater treatment plants (WWTPs), the stochastic nature of influent wastewater and operational and weather conditions cause fluctuations in effluent quality. Data-driven models can forecast effluent quality a few hours ahead as a response to the influent characteristics, providing enough time to adjust system operations and avoid undesired consequences. However, existing data for training models are often incomplete and contain missing values. On the other hand, collecting additional data by installing new sensors is costly. The trade-off between using existing incomplete data and collecting costly new data results in three data challenges faced when developing data-driven WWTP effluent forecasters. These challenges are to determine important variables to be measured, the minimum number of required data instances, and the maximum percentage of tolerable missing values that do not impede the development of an accurate model. As these issues are not discussed in previous studies, in this research, for the first time, a comprehensive analysis is done to provide answers to these challenges. Another issue that arises in all data-driven modeling is how to select an appropriate forecasting model. This paper addresses these issues by first testing nine machine learning models on data collected from three wastewater treatment plants located in Iran, Australia, and Spain. The most accurate forecaster, Bayesian network, was then used to address the articulated challenges. Key variables in forecasting effluent characteristics were flow rate, total suspended solids, electrical conductivity, phosphorus compounds, wastewater temperature, and air temperature. A minimum of 250 samples was needed during the model training to achieve a great reduction in the forecasting error. Moreover, a steep increase in the error was observed should the portion of missing values exceed 10%. The results assist plant managers in estimating the necessary data collection effort to obtain an accurate forecaster, contributing to the quality of the effluent. [Display omitted] • Bayesian network is benchmarked against eight AI models to predict wastewater quality. • Bayesian networks can handle missing value in input data due to probabilistic nature. • Key variables in WWTP quality and quantity were specified. • Input data quality has a huge role in successful development of data-driven models. • The prediction error sharply increases when the portion of missing data exceeds 10%. [ABSTRACT FROM AUTHOR]

This paper presents a new framework for the adaptive reservoir operation considering water quantity and quality objectives. In this framework, using the European Centre for Medium-Range Weather Forecasts (ECMWF) database, the monthly precipitation forecasts, with up to 6-month lead time, are downscaled and bias corrected. The rainfall forecasts are used as inputs to a rainfall-runoff simulation model to predict sub-seasonal inflows to reservoir. The water storage at the end of a short-term planning horizon (e.g. 6 months) is obtained from some probabilistic optimal reservoir storage volume curves, which are developed using a long-term reservoir operation optimization model. The adaptive optimization model is linked with the CE-QUAL-W2 water quality simulation model to assess the quality of outflow from each gate as well as the in-reservoir water quality. At the first of each month, the inflow forecasts for the coming months are updated and operating policies for each gate are revised. To tackle the computational burden of the adaptive simulation-optimization model, it is run using Parallel Cellular Automata with Local Search (PCA-LS) optimization algorithm. To evaluate the applicability and efficiency of the framework, it is applied to the Karkheh dam, which is the largest reservoir in Iran. By comparing the run times of the PCA-LS and the Non-dominated Sorting Genetic Algorithms II (NSGA-II), it is shown that the computational time of PCA-LS is 95 % less than NSGA-II. According to the results, the difference between the objective function of the proposed adaptive optimization model and a perfect model, which uses the observed inflow data, is only 1.68 %. It shows the appropriate accuracy of the adaptive model and justifies using the proposed framework for the adaptive operation of reservoirs considering water quantity and quality objectives. • Adaptive reservoir operation model considering water quantity and quality objectives. • Using ECMWF-based rainfall forecasts with monthly time step and 6-month lead time. • Linking the optimization model with the CE-QUAL-W2 water quality simulation model. • Applying the real-time operation model to the Karkheh river-reservoir system in Iran. • Illustrating the efficiency of the proposed model in a real-world large-scale problem. [ABSTRACT FROM AUTHOR]

Detention reservoirs are employed in urban drainage systems to reduce peak flows downstream of reservoirs. In addition to the volume of detention reservoirs, their operational policies could significantly affect their performance. This paper presents a framework for the real-time coordinated operation of detention reservoirs using deep-learning-based rainfall nowcasting data. Considering the short concentration time of urban basins, the real-time operating policies of urban detention reservoirs should be developed quickly. In the proposed framework, a cellular automata (CA)-based optimization algorithm is linked with the storm water management model (SWMM) to optimize real-time operating policies of gates at the inlets and outlets of detention reservoirs. As CA-based optimization models are not population-based, their computational costs are much less than population-based metaheuristic optimization techniques such as genetic algorithms. To evaluate the applicability and efficiency of the framework, it is applied to the east drainage catchment (EDC) of Tehran metropolitan area in Iran. The results illustrate that the proposed framework could reduce the overflow volume by up to 60%. For complete flood control in the study area, in addition to the real-time operation of detention reservoirs, constructing five tunnels with a total length of 13200 m is recommended. To evaluate the performance of the CA-based optimization model, its results are compared with those obtained from the non-dominated sorting genetic algorithm III (NSGA-III). It is shown that the CA-based model provides similar results with only 5% of the run-time of NSGA-III. A sensitivity analysis is also performed to evaluate the effects of optimization models' parameters on their performance. • Developing a new framework for coordinated operation of urban detention reservoirs. • Using deep-learning-based rainfall nowcasting data for real-time flood management. • Linking a stormwater simulation model with the cellular automata optimization model. • Applying the framework to the east drainage catchment of the Tehran metropolitan area. • Illustrating that the adaptive framework reduces the overflow volume by up to 60%. [ABSTRACT FROM AUTHOR]

Coastal areas are of paramount importance due to their pivotal role in facilitating a wide range of socio-economic activities and providing vital environmental services. These areas, as the meeting points of land and sea, face significant risks of flooding due to the ongoing rise in sea levels caused by climate change. Additionally, they are susceptible to extreme events like king tides and large waves in the future. This paper introduces a framework for estimating the extreme total water level (TWL) by considering the effects of regional sea level rise (RSLR) resulting from a warming climate under RCP 8.5. It also incorporates the contributions of high tides, 100-year storm surge, and 100-year wave setup and run-up. The proposed framework is utilized to evaluate the occurrence of extreme coastal flooding along the Persian Gulf coast of Iran, an area that is home to significant industries in the country. The results offer an estimated increase of RSLR by 0.23 m from 2020 to 2050 considering an ensemble of climate model projections. Extreme wave setup values are estimated to range between 0.19 and 0.66 m, while storm surge is projected to vary from 0.4 to 1.44 m across the studied coastline. These together yield in a projected extreme TWL along the coastline within the range of 3.18 and 3.90 m above the current sea level. This significant increase in sea level could lead to the inundation of approximately 513 km2 of low-lying coastal land, which accounts for about 16% of the studied domain and could pose serious flooding threat to the people and their assets in this region. Finally, relative ranking of flooded zones (i.e., 6 zones) helps determine the areas with higher chance of flood exposure, at which investment in flood mitigation measures should be prioritized. [Display omitted] • Regional sea level rise is quantified using GCMs under RCP 8.5. • Flooding maps are shown under regional SLR, tides, storm surge, wave setup and run-up. • Extreme total water level is expected to inundate important ports in Persian Gulf coast of Iran by 2050. • Research projects that 513 km of low-lying coastal areas will be inundated due to extreme total water level. [ABSTRACT FROM AUTHOR]