The production of high-quality white base, a crucial component in the paper, textile and cleaning products industries, is influenced by various parameters, including the concentration of cationic sulfur trioxide (simply Cat SO₃) which is a surfactant in the process. Elevated Cat SO₃ levels can adversely affect the product's properties, resulting in suboptimal quality and increased costs. This study proposes a data-driven approach to predict Cat SO₃ levels in white base production using data science and statistical data analysis techniques. Traditional methods of Cat SO₃ level prediction often rely on manual monitoring and experience-based adjustments, meaning it can only be calculated after the white base production by a laboratory result, leading to inefficiencies, loss of valuable time and if it is not in the acceptable level, loss of the product . To address this, I employ advanced data science methodologies to develop an accurate predictive model. The proposed model integrates historical process data including laboratory results, and production parameters to forecast Cat SO₃ levels. The methodology encompasses several key steps: data collection and pre-processing, explanatory data analysis (EDA), feature selection, model training, validation, and performance evaluation. Various machine learning algorithms, including regression techniques, are explored to identify the most suitable model for predicting Cat SO₃ levels. Process engineering techniques and engineers support are employed to extract relevant information from the complex and multivariate dataset. The model's effectiveness is evaluated using real production data from Procter and Gamble HDL facility. Performance metrics such as Mean Absolute Error (MAE), Root Mean Square Error (RMSE), and coefficient of determination (R²) are employed to assess the accuracy of predictions. Additionally, the model's robustness and generalization capability are tested against unseen data to ensure its practical utility. Since laboratory analysis also tests more than one results of other value parameters, future Cat SO3 lab analysis results will be available for us to compare the results to new production data. The results indicate that the proposed data-driven approach can replace traditional methods, yielding more accurate and consistent predictions of Cat SO₃ levels in white base production. This research contributes to the optimization of production processes, reducing wastage, and answers the value creation process which facility needs to complete for savings. The application of data science techniques not only explores Cat SO₃ level prediction but also establishes a foundation for further process optimization and automation in the white base manufacturing industry.
The production of high-quality white base, a crucial component in the paper, textile and cleaning products industries, is influenced by various parameters, including the concentration of cationic sulfur trioxide (simply Cat SO₃) which is a surfactant in the process. Elevated Cat SO₃ levels can adversely affect the product's properties, resulting in suboptimal quality and increased costs. This study proposes a data-driven approach to predict Cat SO₃ levels in white base production using data science and statistical data analysis techniques. Traditional methods of Cat SO₃ level prediction often rely on manual monitoring and experience-based adjustments, meaning it can only be calculated after the white base production by a laboratory result, leading to inefficiencies, loss of valuable time and if it is not in the acceptable level, loss of the product . To address this, I employ advanced data science methodologies to develop an accurate predictive model. The proposed model integrates historical process data including laboratory results, and production parameters to forecast Cat SO₃ levels. The methodology encompasses several key steps: data collection and pre-processing, explanatory data analysis (EDA), feature selection, model training, validation, and performance evaluation. Various machine learning algorithms, including regression techniques, are explored to identify the most suitable model for predicting Cat SO₃ levels. Process engineering techniques and engineers support are employed to extract relevant information from the complex and multivariate dataset. The model's effectiveness is evaluated using real production data from Procter and Gamble HDL facility. Performance metrics such as Mean Absolute Error (MAE), Root Mean Square Error (RMSE), and coefficient of determination (R²) are employed to assess the accuracy of predictions. Additionally, the model's robustness and generalization capability are tested against unseen data to ensure its practical utility. Since laboratory analysis also tests more than one results of other value parameters, future Cat SO3 lab analysis results will be available for us to compare the results to new production data. The results indicate that the proposed data-driven approach can replace traditional methods, yielding more accurate and consistent predictions of Cat SO₃ levels in white base production. This research contributes to the optimization of production processes, reducing wastage, and answers the value creation process which facility needs to complete for savings. The application of data science techniques not only explores Cat SO₃ level prediction but also establishes a foundation for further process optimization and automation in the white base manufacturing industry.
Cationic Sulfur Trioxide Level Prediction in White Base Production
ALBUT, SERCAN
2022/2023
Abstract
The production of high-quality white base, a crucial component in the paper, textile and cleaning products industries, is influenced by various parameters, including the concentration of cationic sulfur trioxide (simply Cat SO₃) which is a surfactant in the process. Elevated Cat SO₃ levels can adversely affect the product's properties, resulting in suboptimal quality and increased costs. This study proposes a data-driven approach to predict Cat SO₃ levels in white base production using data science and statistical data analysis techniques. Traditional methods of Cat SO₃ level prediction often rely on manual monitoring and experience-based adjustments, meaning it can only be calculated after the white base production by a laboratory result, leading to inefficiencies, loss of valuable time and if it is not in the acceptable level, loss of the product . To address this, I employ advanced data science methodologies to develop an accurate predictive model. The proposed model integrates historical process data including laboratory results, and production parameters to forecast Cat SO₃ levels. The methodology encompasses several key steps: data collection and pre-processing, explanatory data analysis (EDA), feature selection, model training, validation, and performance evaluation. Various machine learning algorithms, including regression techniques, are explored to identify the most suitable model for predicting Cat SO₃ levels. Process engineering techniques and engineers support are employed to extract relevant information from the complex and multivariate dataset. The model's effectiveness is evaluated using real production data from Procter and Gamble HDL facility. Performance metrics such as Mean Absolute Error (MAE), Root Mean Square Error (RMSE), and coefficient of determination (R²) are employed to assess the accuracy of predictions. Additionally, the model's robustness and generalization capability are tested against unseen data to ensure its practical utility. Since laboratory analysis also tests more than one results of other value parameters, future Cat SO3 lab analysis results will be available for us to compare the results to new production data. The results indicate that the proposed data-driven approach can replace traditional methods, yielding more accurate and consistent predictions of Cat SO₃ levels in white base production. This research contributes to the optimization of production processes, reducing wastage, and answers the value creation process which facility needs to complete for savings. The application of data science techniques not only explores Cat SO₃ level prediction but also establishes a foundation for further process optimization and automation in the white base manufacturing industry.File | Dimensione | Formato | |
---|---|---|---|
Data_Science_MsC_Thesis_UniPD_Sercan_Albut format.pdf
accesso aperto
Dimensione
4.37 MB
Formato
Adobe PDF
|
4.37 MB | Adobe PDF | Visualizza/Apri |
The text of this website © Università degli studi di Padova. Full Text are published under a non-exclusive license. Metadata are under a CC0 License
https://hdl.handle.net/20.500.12608/61372