Feature space monitoring for smart manufacturing via statistics pattern analysis

He, Q. Peter; Wang, Jin; Shah, Devarshi

doi:10.1016/j.compchemeng.2019.04.010

Cited by 26 publications

(17 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The features of vibration data are often used to train ML models to predict fault detection, do predictive maintenance, and for various other diagnostic analyses. He et al (2019) discuss the statistical pattern analysis framework for fault detection and diagnosis for batch and process data. In this approach, instead of process variables, sample-wise and variable-wise statistical features that quantify process characteristics are extracted and used for process monitoring.…”

Section: Robustness Issues Of ML and Ai Modelsmentioning

confidence: 99%

Ensuring the Robustness and Reliability of Data-Driven Knowledge Discovery Models in Production and Manufacturing

Tripathi

Muhr

Manuel

et al. 2021

Front. Artif. Intell.

View full text Add to dashboard Cite

The Cross-Industry Standard Process for Data Mining (CRISP-DM) is a widely accepted framework in production and manufacturing. This data-driven knowledge discovery framework provides an orderly partition of the often complex data mining processes to ensure a practical implementation of data analytics and machine learning models. However, the practical application of robust industry-specific data-driven knowledge discovery models faces multiple data- and model development-related issues. These issues need to be carefully addressed by allowing a flexible, customized and industry-specific knowledge discovery framework. For this reason, extensions of CRISP-DM are needed. In this paper, we provide a detailed review of CRISP-DM and summarize extensions of this model into a novel framework we call Generalized Cross-Industry Standard Process for Data Science (GCRISP-DS). This framework is designed to allow dynamic interactions between different phases to adequately address data- and model-related issues for achieving robustness. Furthermore, it emphasizes also the need for a detailed business understanding and the interdependencies with the developed models and data quality for fulfilling higher business objectives. Overall, such a customizable GCRISP-DS framework provides an enhancement for model improvements and reusability by minimizing robustness-issues.

show abstract

Section: Robustness Issues Of ML and Ai Modelsmentioning

confidence: 99%

Ensuring the Robustness and Reliability of Data-Driven Knowledge Discovery Models in Production and Manufacturing

Tripathi

Muhr

Manuel

et al. 2021

Front. Artif. Intell.

View full text Add to dashboard Cite

show abstract

“…It turns out that the CMDS-PA has successfully detected all the faults effectively in relative to cPCA, whereby cPCA was observed struggling in sustaining the signal detection particularly for fault cases numbers 3, 9, and 15. Perhaps, this outstanding performance can be further utilised as well as expanded for the current use of feature monitoring within the realm of big data environment for smart manufacturing (He et al, 2019;He & Wang, 2018).…”

Section: Resultsmentioning

confidence: 99%

Utilization of Classical Scaling Technique in Sustaining Fault Detection Performance in Process Monitoring

Yunus

Zhang

Al‐Amshawee

2020

JCEIB

View full text Add to dashboard Cite

Multivariate Statistical Process Monitoring (MSPM) fundamentally adopts the conventional Principal Component Analysis (cPCA) as the main platform for data compression. The main challenge though, the association nature of most industrial process variables are highly non-linear. As a result, the risks of applying the conventional approach of MSPM within this context may include sluggish or failed in detection, misinterpretation of signals, incorrect fault diagnosis and also inflexible as well as insensitive to changing of operating modes. In addressing the issue, this paper introduces new sets of monitoring parameters i.e. Sm2, Sr2 and Sr3, which have been derived within the frameworks of Classical Scaling (CMDS) and Procusters Analysis (PA) methods. The overall fault detection performance that applied based on the Tennessee Eastman Process (TEP) cases show that the Sr3 can detect the faults particularly for abnormal events number 3, 9, 15 and 19 in higher rate compared to the cPCA-MSPM system. This proves that the new monitoring statistics work effectively in avoiding missed detection during monitoring which cannot be addressed effectively by the traditional monitoring system.

show abstract

“…Additionally, machine learning techniques have also been improved by incorporating feature engineering ideas to efficiently develop models for large‐scale systems. For example, in Reference 38, it was demonstrated that both machine learning and projection to latent structure (PLS) methods show poor performance on raw vibration signals from a laboratory‐scale water flow system, and therefore, further treatment of raw data sets, such as feature‐based monitoring that can significantly improve model prediction 38‐40 is needed. An implementation of similar machine learning structures can be found for nonlinear systems in References 41–43, and their potential role in Industry 4.0 was also highlighted in (bio)chemical processes 44 .…”

Section: Introductionmentioning

confidence: 99%

Machine learning modeling and predictive control of nonlinear processes using noisy data

Rincón

Liu

et al. 2021

AIChE Journal

View full text Add to dashboard Cite

This work focuses on machine learning modeling and predictive control of nonlinear processes using noisy data. We use long short‐term memory (LSTM) networks with training data from sensor measurements corrupted by two types of noise: Gaussian and non‐Gaussian noise, to train the process model that will be used in a model predictive controller (MPC). We first discuss the LSTM training with noisy data following a Gaussian distribution, and demonstrate that the standard LSTM network is capable of capturing the underlying process dynamic behavior by reducing the impact of noise. Subsequently, given that the standard LSTM performs poorly on a noisy data set from industrial operation (i.e., non‐Gaussian noisy data), we propose an LSTM network using Monte Carlo dropout method to reduce the overfitting to noisy data. Furthermore, an LSTM network using co‐teaching training method is proposed to further improve its approximation performance when noise‐free data from a process model capturing the nominal process state evolution is available. A chemical process example is used throughout the manuscript to illustrate the application of the proposed modeling approaches and demonstrate their open‐ and closed‐loop performance under a Lyapunov‐based model predictive controller with state measurements corrupted by industrial noise.

show abstract

Feature space monitoring for smart manufacturing via statistics pattern analysis

Cited by 26 publications

References 31 publications

Ensuring the Robustness and Reliability of Data-Driven Knowledge Discovery Models in Production and Manufacturing

Ensuring the Robustness and Reliability of Data-Driven Knowledge Discovery Models in Production and Manufacturing

Utilization of Classical Scaling Technique in Sustaining Fault Detection Performance in Process Monitoring

Machine learning modeling and predictive control of nonlinear processes using noisy data

Contact Info

Product

Resources

About