Enhancing sewage flow prediction using an integrated improved SSA-CNN-Transformer-BiLSTM model DOI Creative Commons

Jiawen Ye,

Lei Dai,

HaiYing Wang

et al.

AIMS Mathematics, Journal Year: 2024, Volume and Issue: 9(10), P. 26916 - 26950

Published: Jan. 1, 2024

<p>Accurate prediction of sewage flow is crucial for optimizing treatment processes, cutting down energy consumption, and reducing pollution incidents. Current models, including traditional statistical models machine learning have limited performance when handling nonlinear high-noise data. Although deep excel in time series prediction, they still face challenges such as computational complexity, overfitting, poor practical applications. Accordingly, this study proposed a combined model based on an improved sparrow search algorithm (SSA), convolutional neural network (CNN), transformer, bidirectional long short-term memory (BiLSTM) prediction. Specifically, the CNN part was responsible extracting local features from series, Transformer captured global dependencies using attention mechanism, BiLSTM performed temporal processing features. The SSA optimized model's hyperparameters to improve accuracy generalization capability. validated dataset actual plant. Experimental results showed that introduced mechanism significantly enhanced ability handle data, effectively hyperparameter selection, improving training efficiency. After introducing SSA, CNN, modules, $ {R^{\text{2}}} increased by 0.18744, RMSE (root mean square error) decreased 114.93, MAE (mean absolute 86.67. difference between predicted peak/trough monitored within 3.6% appearance 2.5 minutes away time. By employing multi-model fusion approach, achieved efficient accurate highlighting potential application prospects field treatment.</p>

Language: Английский

How to evaluate uncertainty estimates in machine learning for regression? DOI Creative Commons
Laurens Sluijterman, Eric Cator, Tom Heskes

et al.

Neural Networks, Journal Year: 2024, Volume and Issue: 173, P. 106203 - 106203

Published: Feb. 22, 2024

As neural networks become more popular, the need for accompanying uncertainty estimates increases. There are currently two main approaches to test quality of these estimates. Most methods output a density. They can be compared by evaluating their loglikelihood on set. Other prediction interval directly. These often tested examining fraction points that fall inside corresponding intervals. Intuitively, both seem logical. However, we demonstrate through theoretical arguments and simulations ways have serious flaws. Firstly, cannot disentangle separate components jointly create predictive uncertainty, making it difficult evaluate components. Specifically, confidence reliably estimating performance interval. Secondly, does not allow comparison between directly A better also necessarily guarantee intervals, which is what used in practice. Moreover, current approach intervals has additional We show why testing or single set fundamentally flawed. At best, marginal coverage measured, implicitly averaging out overconfident underconfident predictions. much desirable property pointwise coverage, requiring correct each prediction. practical examples effects result favouring method, based undesirable behaviour Finally, propose simulation-based addresses problems while still allowing easy different methods. This development new quantification

Language: Английский

Citations

12
Artificial intelligence and machine learning for the optimization of pharmaceutical wastewater treatment systems: a review DOI Creative Commons
Voravich Ganthavee, Antoine P. Trzcinski

Environmental Chemistry Letters, Journal Year: 2024, Volume and Issue: 22(5), P. 2293 - 2318

Published: May 21, 2024

Abstract The access to clean and drinkable water is becoming one of the major health issues because most natural waters are now polluted in context rapid industrialization urbanization. Moreover, pollutants such as antibiotics escape conventional wastewater treatments thus discharged ecosystems, requiring advanced techniques for treatment. Here we review use artificial intelligence machine learning optimize pharmaceutical treatment systems, with focus on quality, disinfection, renewable energy, biological treatment, blockchain technology, algorithms, big data, cyber-physical automated smart grid power distribution networks. Artificial allows monitoring contaminants, facilitating data analysis, diagnosing easing autonomous decision-making, predicting process parameters. We discuss advances technical reliability, energy resources management, cyber-resilience, security functionalities, robust multidimensional performance platform distributed consortium, stabilization abnormal fluctuations quality

Language: Английский

Citations

11
Predicting effluent quality parameters for wastewater treatment plant: A machine learning-based methodology DOI

João Vitor Rios Fuck,

Maria Alice Prado Cechinel,

Juliana Neves

et al.

Chemosphere, Journal Year: 2024, Volume and Issue: 352, P. 141472 - 141472

Published: Feb. 19, 2024

Language: Английский

Citations

6
Robust fault detection method based on interval neural networks optimized by ellipsoid bundles DOI Creative Commons
Meng Zhou,

Yinyue Zhang,

Jing Wang

et al.

Automatica, Journal Year: 2025, Volume and Issue: 176, P. 112233 - 112233

Published: March 3, 2025

Language: Английский

Citations

0
Dual purpose of Shapley Additive Explanation (SHAP) in model explanation and feature selection for artificial intelligence-based digital twin of wastewater treatment plant DOI

V Nourani,

Mohammad Ali Dehghan, Aida Hosseini Baghanam

et al.

Journal of Water Process Engineering, Journal Year: 2025, Volume and Issue: 75, P. 107947 - 107947

Published: May 19, 2025

Language: Английский

Citations

0
A novel approach for multivariate time series interval prediction of water quality at wastewater treatment plants DOI Creative Commons
Siyu Liu, Zhaocai Wang, Yanyu Li

et al.

Water Science & Technology, Journal Year: 2024, Volume and Issue: 90(10), P. 2813 - 2841

Published: Nov. 12, 2024

ABSTRACT This study proposes a novel approach for predicting variations in water quality at wastewater treatment plants (WWTPs), which is crucial optimizing process management and pollution control. The model combines convolutional bi-directional gated recursive units (CBGRUs) with adaptive bandwidth kernel function density estimation (ABKDE) to address the challenge of multivariate time series interval prediction WWTP quality. Initially, wavelet transform (WT) was employed smooth data, reducing noise fluctuations. Linear correlation coefficient (CC) non-linear mutual information (MI) techniques were then utilized select input variables. CBGRU applied capture temporal correlations series, integrating Multiple Heads Attention (MHA) mechanism enhance model's ability comprehend complex relationships within data. ABKDE employed, supplemented by bootstrap establish upper lower bounds intervals. Ablation experiments comparative analyses benchmark models confirmed superior performance point prediction, analysis forecast period, fluctuation detection Also, this verifies broad applicability robustness anomalous contributes significantly improved effluent efficiency control WWTPs.

Language: Английский

Citations

2
Analyzing the impact of artificial intelligence on operational efficiency in wastewater treatment: a comprehensive neutrosophic AHP-based SWOT analysis DOI
Selin Yalçın, Ertuğrul Ayyıldız

Environmental Science and Pollution Research, Journal Year: 2024, Volume and Issue: 31(38), P. 51000 - 51024

Published: Aug. 6, 2024

Language: Английский

Citations

1
The AI Cleanse: Revolutionizing the Future of Wastewater Treatment with AI and Machine Learning DOI

R. Sanjeevi,

Prashantkumar B. Sathvara, Sandeep Tripathi

et al.

Springer water, Journal Year: 2024, Volume and Issue: unknown, P. 245 - 263

Published: Jan. 1, 2024

Language: Английский

Citations

1
Particle Swarm Training of a Neural Network for the Lower Upper Bound Estimation of the Prediction Intervals of Time Series DOI Creative Commons
Alexander Gusev, Alexander Chervyakov,

Anna Alexeenko

et al.

Mathematics, Journal Year: 2023, Volume and Issue: 11(20), P. 4342 - 4342

Published: Oct. 19, 2023

Many time series forecasting applications use ranges rather than point forecasts. Producing forecasts in the form of Prediction Intervals (PIs) is natural, since intervals are an important component many mathematical models. The LUBE (Lower Upper Bound Estimation) method aimed at finding based on solving optimization problems taking into account interval width and coverage. Using Particle Swarm Training simple neural network, we look for a solution to problem Coverage Width-Based Criterion (CWC), which exponential convolution conflicting criteria PICP (Prediction Interval Probability) PINRW Normalized Root-mean-square Width). Based concept Pareto compromise, it introduced as front space specified criteria. compromise constructed relationship between found problem. data under consideration financial MOEX closing prices. Our findings reveal that relatively comprising eight neurons their corresponding 26 parameters (weights neuron connections signal biases), sufficient yield reliable PIs investigated series. novelty our approach lies network structure (containing fewer 100 parameters) construct Additionally, offer experimental construction frontier, formed by

Language: Английский

Citations

1
Exploring stochastic differential equation for analyzing uncertainty in wastewater treatment plant-activated sludge modeling DOI Creative Commons

Reza Shahidi Zonouz,

Vahid Nourani, Mina Sayyah-Fard

et al.

AQUA - Water Infrastructure Ecosystems and Society, Journal Year: 2024, Volume and Issue: 73(3), P. 520 - 537

Published: Feb. 10, 2024

Abstract The management of wastewater treatment plant (WWTP) and the assessment uncertainty in its design are crucial from an environmental engineering perspective. One key mechanisms WWTP operation is activated sludge, which related to biological oxygen demand (BOD) parameter. In modeling BOD, conventional approach utilizing ordinary differential equations (ODEs) fails incorporate stochastic nature this parameter, leading a considerable level WWTP. To address issue, study proposes model that utilizes (SDEs) instead ODE simulate BOD activities microorganisms flow rate (Q). SDEs integral It̂o solved using Euler–Maruyama method for period 15 sequential days timespan 2019–2020 Tabriz City. SDE improves facilities by identifying uncertainties enhancing reliability. This, turn, increases reliability technical structures within performance system. By considering randomness problem, proposed significantly results, with enhancement 11.47 10.11% Q models, respectively.

Language: Английский

Citations

0