Data-Driven Hybrid Model for Forecasting Wastewater Influent Loads Based on Multimodal and Ensemble Deep Learning

Heo, SungKu; Nam, KiJeon; Loy-Benitez, Jorge

doi:10.1109/tii.2020.3039272

Cited by 43 publications

(10 citation statements)

References 28 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…From these models the MBNA model is observed to be most effective in terms of quality and sludge levels. While, in terms of computational complexity PD (Deng et al 2020), AOPO (Deng et al 2020), SVI (Khan et al 2018), IMCDR (Elsahwi et al 2019), FLSWTP (Lamas & Giacaglia 2021), and MA (Deng et al 2020) are the most effective, but, MEDL (Heo et al 2021), FLSWTP (Lamas & Giacaglia 2021), MBNA (Cheriyamundath & Vavilala 2021), SPIC (Tung et al 2018), MWH (Karlsson et al 2019), and NP (Cheriyamundath & Vavilala 2021) have the lowest treatment delay. This is because the later models utilize chemical reactions in order to speed-up the treatment process.…”

Section: Discussionmentioning

confidence: 99%

“…Efficiency of such processes can be improved via the addition of influent load forecasting in wastewater, and then scheduling treatment activities accordingly. The work in Heo et al (2021) proposes such a model, wherein authors have discussed use of multimodal and ensemble deep learning (MEDL) for prediction of wastewater output from systems, which assists in scheduling treatment activities. Due to this scheduling, delay needed for treatment reduces drastically, and availability of usable water is improved.…”

Section: Literature Reviewmentioning

confidence: 99%

See 1 more Smart Citation

Pragmatic analysis of wastewater treatment methods from a statistical perspective

Pande

Bhagat

2022

Water Practice and Technology

View full text Add to dashboard Cite

Wastewater treatment is an environmental issue of utmost importance. Various models and architectures for wastewater treatment exist, each having its unique characteristics. In this study, an application-specific review of various treatment models is performed. Extensions to existing treatment models are discussed for improvement in process performance. The treatment models are compared statistically based on their performance metrics, namely the quality of treated water (Q), sludge percentage at output (SL), complexity of treatment (C), time needed for treatment (T), and deployment cost (DC). A novel parameter, model rank, is proposed that combines all the performance metrics into a single number so that the treatment models can be analyzed effectively. Results show that advanced oxidation processes with ozone treatment (AOPO), Kernel principal components analysis-based one-class support vector machine (KPCA SVM), electrochemical processes (EPs), membrane and absorption (MA), Nondominated Sorting Genetic Algorithm-based Optimal Controller (NSGAOC), and wet-type nonthermal plasma reactor (WTNPR) models have a rank above 3.5. The AOPO model has the highest model rank of 3.85 and thus has better overall performance than others. This study might aid major stakeholders in waste treatment industry including researchers in selecting the appropriate wastewater treatment method according to their requirements.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Literature Reviewmentioning

confidence: 99%

Pragmatic analysis of wastewater treatment methods from a statistical perspective

Pande

Bhagat

2022

Water Practice and Technology

View full text Add to dashboard Cite

show abstract

“…10 More recently, deep learning-based methods have been developed for monitoring multimode process. For example, a multimodal and ensemble-based deep learning method 11 was proposed to deal with highly nonlinear characteristics. Similarly, a multi-branch deep neural network 12 was proposed to extract and fuse features from different process modes.…”

Section: Introductionmentioning

confidence: 99%

Structured discriminative Gaussian graph learning for multimode process monitoring

Wang,

Liu,

Zhang

et al. 2024

Journal of Chemometrics

View full text Add to dashboard Cite

Aiming at the actual industrial process background that different modes share the same system configurations and control structure, this article proposes a novel structured discriminant Gaussian graph learning for multimode process monitoring. The proposed method considers not only the sparsity of graph model but also the measurement of data variation based on a mismatched graph and the common node support between different graphical structures. The objective function involves two sets of regularization terms: the trace terms for mismatched measurements and the ‐norm imposed on the union of decomposed graph matrices. Due to the introduced mismatched trace terms, the cost of matching the data points and graph models that have inconsistent class labels can be expanded, which brings more discrimination for the graph‐based mode identification. While the common structure extracted by the ‐norm forces the estimated graph models to have structural similarities, thus alleviating the negative influence caused by graph discrimination. Once a relatively accurate and discriminative reference graph model is obtained, the downstream test graph learning and analysis can be conducted online by employing the moving window techniques. By comparing the matched and mismatched graph‐based measurements, the process mode can be identified correctly and stably. To grasp the abnormal process changes, the ‐norm for the row sparsity is again applied to the graph difference matrices, the sensitive monitoring statistics and the fault isolation results can be obtained effectively. All the optimization problems in this paper can be solved using the alternating direction multiplier (ADMM) algorithm. The effectiveness of our proposed approach is illustrated by the application to a real blast furnace iron‐making production process.

show abstract

“…Cheng et al [27] used an adaptive network-based fuzzy inference system (ANFIS) to predict the influent characteristics of wastewater treatment. Heo et al [28] established a hybrid influent forecasting model which was based on multimodal and ensemble-based deep learning (ME-DeepL). This model exhibited applications in fluctuating influent loads, as it can capture the informative features and temporal patterns.…”

Section: Introductionmentioning

confidence: 99%

The Impact of Wastewater Quality and Flow Characteristics on H2S Emissions Generation: Statistical Correlations and an Artificial Neural Network Model

Sherief

Hassan

2022

Water

View full text Add to dashboard Cite

Hydrogen sulfide (H2S) is a naturally occurring, highly toxic gas that is formed from the decomposition of sulfur compounds. H2S is a common source of concrete and metal corrosion that results in huge economic losses in wastewater collection and treatment plants. Hence, it is necessary analyze H2S generation and emission. H2S concentrations were measured at the Al-Saad wastewater treatment plant in United Arab Emirates. Wastewater samples were collected, and water quality parameters were characterized in the laboratory. Simultaneously, flow characteristics, humidity, headspace airflow, and temperature were measured onsite. A neural network model to predict H2S emissions was formulated using significant parameters. It was observed that flowrate, velocity, sulfate, and total sulfur had a similar cyclic pattern throughout the sampling events. The temperature, humidity, total sulfur, and depth of wastewater were identified as the most important parameters influencing H2S emissions through correlation analysis. The neural model validation and testing had an R value of 0.9. The training had an R value of 0.8. The model provided an accuracy of 80% for the prediction of H2S concentration in wastewater treatment plants. The accuracy can be improved by increasing data. The model is limited to its applicability in the prediction of H2S emissions under conditions similar to the inlet of a wastewater treatment plant.

show abstract

Data-Driven Hybrid Model for Forecasting Wastewater Influent Loads Based on Multimodal and Ensemble Deep Learning

Cited by 43 publications

References 28 publications

Pragmatic analysis of wastewater treatment methods from a statistical perspective

Pragmatic analysis of wastewater treatment methods from a statistical perspective

Structured discriminative Gaussian graph learning for multimode process monitoring

The Impact of Wastewater Quality and Flow Characteristics on H2S Emissions Generation: Statistical Correlations and an Artificial Neural Network Model

Contact Info

Product

Resources

About