Water quality soft-sensor prediction in anaerobic process using deep neural network optimized by Tree-structured Parzen Estimator

● Hybrid deep-learning model is proposed for water quality prediction. ● Tree-structured Parzen Estimator is employed to optimize the neural network. ● Developed model performs well in accuracy and uncertainty. ● Usage of the proposed model can reduce carbon emission and energy consumption. Anaerobic process is regarded as a green and sustainable process due to low carbon emission and minimal energy consumption in wastewater treatment plants (WWTPs). However, some water quality metrics are not measurable in real time, thus influencing the judgment of the operators and may increase energy consumption and carbon emission. One of the solutions is using a soft-sensor prediction technique. This article introduces a water quality soft-sensor prediction method based on Bidirectional Gated Recurrent Unit (BiGRU) combined with Gaussian Progress Regression (GPR) optimized by Tree-structured Parzen Estimator (TPE). TPE automatically optimizes the hyperparameters of BiGRU, and BiGRU is trained to obtain the point prediction with GPR for the interval prediction. Then, a case study applying this prediction method for an actual anaerobic process (2500 m³/d) is carried out. Results show that TPE effectively optimizes the hyperparameters of BiGRU. For point prediction of COD_eff and biogas yield, R² values of BiGRU, which are 0.973 and 0.939, respectively, are increased by 1.03%–7.61% and 1.28%–10.33%, compared with those of other models, and the valid prediction interval can be obtained. Besides, the proposed model is assessed as a reliable model for anaerobic process through the probability prediction and reliable evaluation. It is expected to provide high accuracy and reliable water quality prediction to offer basis for operators in WWTPs to control the reactor and minimize carbon emission and energy consumption.     

Distribution and assessment of heavy metal toxicity in sediment cores from Bizerte Lagoon,Tunisia

To examine the state of pollution of Bizerte Lagoon which is exposed to intense anthropogenic pressure, two sediment cores were taken at two sites, one undergoes the dual effects of both marine waters arriving from the Mediterranean Sea through the Channel, and also of freshwater from the Tinja River; the other core is located at the center of the lagoon where water depth is maximal (12 m). Heavy metal concentrations in the two cores were assessed, with calculated enrichment factors and geo-accumulation indexes. Core sediments were also studied for chemical speciation and their monosulfide contents were measured. Results from enrichment factors and geo-accumulation indexes show an accumulation of Cd, Zn, Cr, and Pb, while chemical speciation revealed a risk only from Cd and Mn. Comparison of sequential extraction values with those of acid volatile sulfides revealed that non-toxic effects may be caused by any of the studied metals in the sediment.     

The impact of Local Authorities’ interventions on household waste collection: A case study approach using time series modelling

At a local Government level there have been many interventions and changes made to household waste collection services to meet new regulatory requirements. These changes include separate collection of recyclable and organic materials. This paper has used a time series model to quantify the success of interventions introduced by a LA.The case study was a medium sized UK LA, Charnwood Borough Council (CBC), the research analyses monthly data of quantities of recyclates, garden waste for composting and residual waste for landfill disposal. The time series model was validated with a five year data set and used to measure the impacts of the various changes to identify which intervention was the most successful, while controlling for season and number of working days. The results show the interventions analysed both had abrupt and permanent positive impacts on the yield of recyclable materials, and a corresponding negative impact on the residual waste.The model could be added to the National data base to help LAs to compare interventions and to understand which schemes encourage householder participation and improve recycling performance.     

Characteristics of granular sludge developed in an upflow anaerobic sludge fixed-film bioreactor treating palm oil mill effluent.

In the present study, characteristics of the granular sludge (including physical characteristics under stable conditions and process shocks arising from suspended solid overload, soluble organic overload, and high temperature; biological activity; and sludge kinetic evaluation in a batch experiment) developed in an upflow anaerobic sludge blanket fixed-film reactor for palm oil mill effluent (POME) treatment was investigated. The main aim of this work was to provide suitable understanding of POME anaerobic digestion using such a granular sludge reactor, particularly with respect to granule structure at various operating conditions. The morphological changes in granular sludge resulting from various operational conditions was studied using scanning electron microscopy and transmission electron microscopy images. It was shown that the developed granules consisted of densely packed rod- (Methanosaeta-like microorganism; predominant) and cocci- (Methanosarsina) shaped microorganisms. Methanosaeta aggregates functioned as nucleation centers that initiated granule development of POME-degrading granules. Under the suspended solid overload condition, most of the granules were covered with a thin layer of fiberlike suspended solids, so that the granule color changed to brown and the sludge volume index also increased to 24.5 from 12 to 15 mL/g, which caused a large amount of sludge washout. Some of the granules were disintegrated because of an acidified environment, which originated from acidogenesis of high influent organic load (29 g chemical oxygen demand [COD]/L d). At 60 degrees C, the rate of biomass washout increased, as a result of disintegration of the outer layer of the granules. In the biological activity test, approximately 95% COD removal was achieved within 72 hours, with an initial COD removal rate of 3.5 g COD/L d. During POME digestion, 275 mg calcium carbonate/L bicarbonate alkalinity was produced per 1000 mg COD(removed)/ L. A consecutive reaction kinetic model was used to simulate the data obtained from the sludge activity in the batch experiment. The mathematical model gave a good fit with the experimental results (R2 > 0.93). The slowest step was modeled to be the acidification step, with a rate constant between 0.015 and 0.083 hours(-1), while the rate constant for the methanogenic step was obtained to be between 0.218 and 0.361 hours(-1).     

Natural radioactivity in farm soil and phosphate fertilizer and its environmental implications in Qena governorate, Upper Egypt

Samples of phosphate fertilizers and farm soils, taken to a depth of up to 30 cm in cultivated land, were collected over the Qena governorate, Upper Egypt. Activity concentration of background radionuclides such as (226)Ra, (232)Th and (40)K of these samples were determined by gamma-ray spectrometry. The results show that these radionuclides were present in concentrations of 366+/-10.5, 66.7+/-7.3 and 4+/-2.6 Bq/kg for phosphate fertilizers. For farm soil and Nile island's soil the corresponding values were 13.7+/-7, 12.3+/-4.6, 1233+/-646 and 11.9+/-6.7, 10.5+/-6.1, 1636+/-417 Bq/kg, respectively. The radium equivalent activity (Ra(eq)), the representative level index, I(gamma r), and absorbed dose in air for all samples were calculated. The data were discussed and compared with those given in the literature.     

Identifying crash propensity using specific traffic speed conditions

INTRODUCTION: In spite of recent advances in traffic surveillance technology and ever-growing concern over traffic safety, there have been very few research efforts establishing links between real-time traffic flow parameters and crash occurrence. This study aims at identifying patterns in the freeway loop detector data that potentially precede traffic crashes. METHOD: The proposed solution essentially involves classification of traffic speed patterns emerging from the loop detector data. Historical crash and loop detector data from the Interstate-4 corridor in the Orlando metropolitan area were used for this study. Traffic speed data from sensors embedded in the pavement (i.e., loop detector stations) to measure characteristics of the traffic flow were collected for both crash and non-crash conditions. Bayesian classifier based methodology, probabilistic neural network (PNN), was then used to classify these data as belonging to either crashes or non-crashes. PNN is a neural network implementation of well-known Bayesian-Parzen classifier. With its superb mathematical credentials, the PNN trains much faster than multilayer feed forward networks. The inputs to final classification model, selected from various candidate models, were logarithms of the coefficient of variation in speed obtained from three stations, namely, station of the crash (i.e., station nearest to the crash location) and two stations immediately preceding it in the upstream direction (measured in 5 minute time slices of 10-15 minutes prior to the crash time). RESULTS: The results showed that at least 70% of the crashes on the evaluation dataset could be identified using the classifiers developed in this paper.