{"title":"人工神经网络优化吸附动力学机理","authors":"Djebbar Mustapha , Thenia Ahmed","doi":"10.1016/j.cdc.2023.101072","DOIUrl":null,"url":null,"abstract":"<div><h3>Abstract</h3><p>Salicylic acid removal by clay was investigated. Isotherms and adsorption kinetics have been optimized to calculate retentions. Experiments for determining the adsorption isotherms were reviewed, including the effect of pH variation and initial salicylic acid concentration.</p><p>The results were modeled using the artificial neural network (ANN) and the pseudo-first and second order. We used MATLAB software to determine the test, validation, and overall regression value.</p><p>The experimental results of the global reaction have non-significant regression coefficients which are adjustable to the pseudo-second order. One of the crucial tasks in the creation of the AAN model is optimizing each of these variables. Salicylic acid adsorption tests at different initial concentrations on natural and treated clay were carried out for 60 min. Mean square error (MSE) data were utilized to determine the ideal number of neurons in the current study, which optimized the hidden layer's number of neurons to 15 for each layer. The ANN model optimized above matches salicylic acid adsorption on Clay better than the Pseudo Second-order, as seen by all the regression values being near to 1.</p></div>","PeriodicalId":269,"journal":{"name":"Chemical Data Collections","volume":null,"pages":null},"PeriodicalIF":2.2180,"publicationDate":"2023-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"1","resultStr":"{\"title\":\"Adsorption kinetics mechanism optimized by artificial neural network\",\"authors\":\"Djebbar Mustapha , Thenia Ahmed\",\"doi\":\"10.1016/j.cdc.2023.101072\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<div><h3>Abstract</h3><p>Salicylic acid removal by clay was investigated. Isotherms and adsorption kinetics have been optimized to calculate retentions. Experiments for determining the adsorption isotherms were reviewed, including the effect of pH variation and initial salicylic acid concentration.</p><p>The results were modeled using the artificial neural network (ANN) and the pseudo-first and second order. We used MATLAB software to determine the test, validation, and overall regression value.</p><p>The experimental results of the global reaction have non-significant regression coefficients which are adjustable to the pseudo-second order. One of the crucial tasks in the creation of the AAN model is optimizing each of these variables. Salicylic acid adsorption tests at different initial concentrations on natural and treated clay were carried out for 60 min. Mean square error (MSE) data were utilized to determine the ideal number of neurons in the current study, which optimized the hidden layer's number of neurons to 15 for each layer. The ANN model optimized above matches salicylic acid adsorption on Clay better than the Pseudo Second-order, as seen by all the regression values being near to 1.</p></div>\",\"PeriodicalId\":269,\"journal\":{\"name\":\"Chemical Data Collections\",\"volume\":null,\"pages\":null},\"PeriodicalIF\":2.2180,\"publicationDate\":\"2023-10-01\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"1\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Chemical Data Collections\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://www.sciencedirect.com/science/article/pii/S2405830023000836\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q2\",\"JCRName\":\"Chemistry\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Chemical Data Collections","FirstCategoryId":"1085","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S2405830023000836","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q2","JCRName":"Chemistry","Score":null,"Total":0}
Adsorption kinetics mechanism optimized by artificial neural network
Abstract
Salicylic acid removal by clay was investigated. Isotherms and adsorption kinetics have been optimized to calculate retentions. Experiments for determining the adsorption isotherms were reviewed, including the effect of pH variation and initial salicylic acid concentration.
The results were modeled using the artificial neural network (ANN) and the pseudo-first and second order. We used MATLAB software to determine the test, validation, and overall regression value.
The experimental results of the global reaction have non-significant regression coefficients which are adjustable to the pseudo-second order. One of the crucial tasks in the creation of the AAN model is optimizing each of these variables. Salicylic acid adsorption tests at different initial concentrations on natural and treated clay were carried out for 60 min. Mean square error (MSE) data were utilized to determine the ideal number of neurons in the current study, which optimized the hidden layer's number of neurons to 15 for each layer. The ANN model optimized above matches salicylic acid adsorption on Clay better than the Pseudo Second-order, as seen by all the regression values being near to 1.
期刊介绍：
Chemical Data Collections (CDC) provides a publication outlet for the increasing need to make research material and data easy to share and re-use. Publication of research data with CDC will allow scientists to: -Make their data easy to find and access -Benefit from the fast publication process -Contribute to proper data citation and attribution -Publish their intermediate and null/negative results -Receive recognition for the work that does not fit traditional article format. The research data will be published as ''data articles'' that support fast and easy submission and quick peer-review processes. Data articles introduced by CDC are short self-contained publications about research materials and data. They must provide the scientific context of the described work and contain the following elements: a title, list of authors (plus affiliations), abstract, keywords, graphical abstract, metadata table, main text and at least three references. The journal welcomes submissions focusing on (but not limited to) the following categories of research output: spectral data, syntheses, crystallographic data, computational simulations, molecular dynamics and models, physicochemical data, etc.
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号
