A Hybrid Process Monitoring Strategy for Steel Making Shop

  • Ashish Kumar
  • Anupam DasEmail author
  • Swarnambuj Suman
Conference paper
Part of the Lecture Notes in Mechanical Engineering book series (LNME)


The article deals with the development of a process monitoring strategy for a Steel Making Shop (SMS) involving an ensemble of statistical and AI techniques. The monitoring strategy being devised was employed primarily to demonstrate the monitoring of nonlinear processes. The monitoring strategy was based on neural network fitting model and Hotelling T2 control chart. Data pertaining to process and feedstock characteristics of Steel Making Shop (SMS) was considered for checking the efficacy of the monitoring strategy being devised. The neural network fitting model is used for partial or full transformation of the nonlinear data into linear data. Thereafter Hotelling T2 chart was employed on the transformed data for monitoring of the process. The test of nonlinearity of the data involved pairwise comparison of any two characteristics by plotting them in a fitted line plot and computing the model fit value which is indicative of the level of linearity. The hybrid strategy involving the neural network model and Hotelling T2 square chart thus devised was able to monitor the process and detect out of control observation correctly.


Neural network fitting model Hotelling T2 chart Steel making shop (SMS) 


  1. 1.
    Montgomery DC (2004) Introduction to statistical quality control, 4th edn. Wiley, New YorkzbMATHGoogle Scholar
  2. 2.
    MacGregor JF, Kourti T (1995) Statistical process control of multivariate processes. Control Eng Pract 3:403–414CrossRefGoogle Scholar
  3. 3.
    Akeem AA, Abubakar Y, Osebekwin A (2015) Hotelling’s T2 decomposition approach for five process characteristics in a multivariate statistical process control. Am J Theor Appl Stat 4(6):432–437CrossRefGoogle Scholar
  4. 4.
    Mostajeran A, Iranpanah N, Noorossana R (2016) A new bootstrap based algorithm for Hotelling’s T2 multivariate control chart. J Sci Islamic Repub Iran 27(3):269–278MathSciNetGoogle Scholar
  5. 5.
    Zhou L, Junghui C, Beiping H, Zhihuan S (2018) Multi-grade principal component analysis for fault detection with multiple production grades. Chemometr Intell Lab Syst 175(4):20–29CrossRefGoogle Scholar
  6. 6.
    Suman S, Das A (2019) Stratified statistical monitoring strategy for a multi-product manufacturing facility with early detection approach. Comput Ind Eng 130:551–564CrossRefGoogle Scholar
  7. 7.
    Wang H, Liu Q, Tu Y (2005) Interpretation of partial least-squares regression models with VARIMAX rotation. Comput Stat Data Anal 48(1):207–219MathSciNetCrossRefGoogle Scholar
  8. 8.
    Nguyen VH, Golinval JC (2010) Fault detection based on kernel principal component analysis. Eng Struct 32:3683–3691CrossRefGoogle Scholar
  9. 9.
    Zhiqiang G, Muguang Z, Zhihuan S (2010) Nonlinear process monitoring based on linear subspace and bayesian inference. J Process Control 20(5):676–688CrossRefGoogle Scholar
  10. 10.
    Hsieh WW (2001) Nonlinear principal component analysis by neural networks. Tellus A 53(5):599–615CrossRefGoogle Scholar
  11. 11.
    Stern HS (1996) Neural networks in applied statistics. Technometrics 38(3):205–214MathSciNetCrossRefGoogle Scholar
  12. 12.
    Jiang L, Song Z, Ge Z, Chen J (2017) Robust self-supervised model and its application for fault detection. Ind Eng Chem Res 56(26):7503–7515CrossRefGoogle Scholar

Copyright information

© Springer Nature Singapore Pte Ltd. 2021

Authors and Affiliations

  1. 1.National Institute of Technology PatnaPatnaIndia

Personalised recommendations