Defect Data Modeling and Analysis for Improving Product Quality and Productivity in Steel Industry

Zhang, Xinmin; Kano, Manabu; Tani, Masahiro; Mori, Junichi; Harada, Kohhei

doi:10.1016/b978-0-444-64241-7.50367-0

Cited by 6 publications

(4 citation statements)

References 5 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Specifically, subsurface inclusion defects refer to irregular and discontinuous slag chunks embedded in the surface or 2∼10 mm under the surface. It can cause serious defects in the resultant hot rolling or cold rolling products, increasing the defective index, the probability of breakout accidents, and the complexity of the hot rolling process [ 3 , 4 , 5 ].…”

Section: Introductionmentioning

confidence: 99%

Real-Time Forecasting of Subsurface Inclusion Defects for Continuous Casting Slabs: A Data-Driven Comparative Study

Wei

Song

2023

Sensors

View full text Add to dashboard Cite

Subsurface inclusions are one of the most common defects that affect the inner quality of continuous casting slabs. This increases the defects in the final products and increases the complexity of the hot charge rolling process and may even cause breakout accidents. The defects are, however, hard to detect online by traditional mechanism-model-based and physics-based methods. In the present paper, a comparative study is carried out based on data-driven methods, which are only sporadically discussed in the literature. As a further contribution, a scatter-regularized kernel discriminative least squares (SR-KDLS) model and a stacked defect-related autoencoder back propagation neural network (SDAE-BPNN) model are developed to improve the forecasting performance. The scatter-regularized kernel discriminative least squares is designed as a coherent framework to directly provide forecasting information instead of low-dimensional embeddings. The stacked defect-related autoencoder back propagation neural network extracts deep defect-related features layer by layer for a higher feasibility and accuracy. The feasibility and efficiency of the data-driven methods are demonstrated through case studies based on a real-life continuous casting process, where the imbalance degree drastically vary in different categories, showing that the defects are timely (within 0.01 ms) and accurately forecasted. Moreover, experiments illustrate the merits of the developed scatter-regularized kernel discriminative least squares and stacked defect-related autoencoder back propagation neural network methods regarding the computational burden; the F1 scores of the developed methods are clearly higher than common methods.

show abstract

Section: Introductionmentioning

confidence: 99%

Real-Time Forecasting of Subsurface Inclusion Defects for Continuous Casting Slabs: A Data-Driven Comparative Study

Wei

Song

2023

Sensors

View full text Add to dashboard Cite

show abstract

“…To remain competitive notably requires continuous innovation and cost reduction strategies during the production of dairy-based products (Kong, Yang & Xu, 2019;Magnus-son & Berggren, 2017). In terms of improving a product's quality and productivity, reducing product defects may be considered as a strategy (Zhang, Kano, Tani, Mori & Harada, 2018). According to Ahmad and Ginantaka (2018), product defect categories in commercially sterilized bottled milk are leak, blown, nonstandard label, no coding, dented bottle, narrow seal, folded, wrinkle and overheat.…”

Section: Introductionmentioning

confidence: 99%

Industrial Practice for Reducing Defective Sterile Milk Products Produced Using Overpressure Rotary Retorts

Pamuji

Purnomo

Sitanggang

2021

IJFS

View full text Add to dashboard Cite

Indonesian consumers are fond of commercially sterilized milk as indicated by increasing product sales. High demand for products intensifies the need to increase productivity, generally achieved by minimizing product defects. This study aimed to reduce the number of defects in commercially sterilized milk produced using overpressure rotary retorts. Based on Pareto analysis, the percentage of defective products was 5.14% of which 2.37% were dented bottles. A cause-effect diagram (Ishikawa Diagram) was used to find the root cause of dented bottles. The pressure difference between the retort chamber (external pressure) and inside the product packaging (internal pressure), and the number of bottles stacked inside the retort basket (bottle density) were found as major factors for causing dented bottles. The internal pressure was 1.20 bar higher than the external pressure. By reducing the pressure difference to 0.40 bar, the percentage of dented bottles could be reduced to 0.79%. Applying the low-est bottle density (73% of the retort basket area occupied by bottles) during the sterilization process could decrease the number of dented bottles, however, it also increased the appearance of striped lids. The best conditions for sterilization (pressure difference = 0.40 bar; number of bottles/basket = 1938 bottles) which were used in the three-month full-scale production trial reduced the percentage of defective products from 5.14% to 2.24% of which 0.76% were dented bottles. Setting the retort pressure at 2.80 bar could avoid 52,920 defective bottles of commercially sterilized products per month.

show abstract

“…Unlike the first-principles models which require the underlying chemical and physical principles of the modeling process, data-driven models are developed directly from the data. In recent years, data-driven soft-sensors have attracted growing attention and achieved successful applications in various areas. − Although data-driven soft-sensors can be constructed in a variety of ways, the focus of this study is on the just-in-time (JIT) learning method, since it can handle process nonlinearity and time-varying characteristics . As illustrated in the questionnaire survey in Japan, model maintenance is considered to be the most important issue associated with soft-sensors, as the prediction performance of soft-sensors may be degraded due to changes in process characteristics.…”

Section: Introductionmentioning

confidence: 99%

Optimal Weighting Distance-Based Similarity for Locally Weighted PLS Modeling

Zhang

Kano

Song

2020

Ind. Eng. Chem. Res.

Self Cite

View full text Add to dashboard Cite

Real-time prediction of product quality or other key performance indicators is critical to ensuring high-quality products and increasing economic profit. In this study, a new locally weighted partial least-squares (LW-PLS) method using optimal weighting distance-based similarity, denoted as OLW-PLS, is proposed. OLW-PLS is a nonlinear just-in-time modeling method, which can handle process collinearity, nonlinearity, and time-varying characteristics. In OLW-PLS, a weighted PLS regression model is constructed based on the optimal weighting distance-based similarity, which considers variable interactions and nonlinear dependencies between the input variables and the output in an optimal manner. The feasibility and effectiveness of the proposed OLW-PLS method were validated through its applications to a numerical example, an industrial ethylene fractionation process, and a pharmaceutical process. The application results have demonstrated that OLW-PLS has superior prediction performance than the conventional PLS, LW-PLS, and CbLW-PLS.

show abstract

Defect Data Modeling and Analysis for Improving Product Quality and Productivity in Steel Industry

Cited by 6 publications

References 5 publications

Real-Time Forecasting of Subsurface Inclusion Defects for Continuous Casting Slabs: A Data-Driven Comparative Study

Real-Time Forecasting of Subsurface Inclusion Defects for Continuous Casting Slabs: A Data-Driven Comparative Study

Industrial Practice for Reducing Defective Sterile Milk Products Produced Using Overpressure Rotary Retorts

Optimal Weighting Distance-Based Similarity for Locally Weighted PLS Modeling

Contact Info

Product

Resources

About