This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.
IEEE TRANSACTIONS ON CONTROL SYSTEMS TECHNOLOGY 1
Incipient Fault Detection for Multiphase Batch
Processes With Limited Batches
Shumei Zhang, Chunhui Zhao , Senior Member, IEEE, and Furong Gao
Abstract—Sufficient batches are in general, required for fault
detection of batch processes. However, sometimes, it is difficult
and may be impractical to conduct multiple cycles and wait until
enough batches are available. Without batch-wise normalization
step, the nonstationary cannot be efficiently removed by time-wise
normalization. The mean values of the nonstationary variables
are still time varying and the interval of fault-free data is
very wide in each phase. Therefore, the incipient fault, which
has a small magnitude including early changing and the slow
developing, may be buried by nonstationary trends resulting in
low fault detection rate. To address the above issue, a two-layer
fault detection method is proposed to detect the incipient fault
for multiphase batch processes with limited batches. First, a con-
current variable separation strategy is proposed to distinguish
nonstationary variables from stationary variables for multiple
batches in each phase. Second, a two-layer fault detection model
is constructed to detect the incipient fault. Cointegration analysis-
based fault detection model is built to investigate the relationship
between nonstationary variables, which can effectively distinguish
the incipient fault from the normal trend of the nonstationary
variables. Principal component analysis is adopted to describe
the correlation of stationary variables. Afterward, a total fault
detection model is constructed to monitor the relation between
nonstationary variables and stationary variables. To illustrate the
feasibility and effectiveness, the proposed algorithm is applied
to two multiphase batch processes including fed-batch penicillin
fermentation process and semiconductor etch process.
Index Terms—Concurrent variable separation, incipient fault,
limited batches, multiphase batch processes, two-layer fault
detection strategy.
I. INTRODUCTION
B
ATCH and semi-batch processes have been playing
a significant role in modern industrial processes such
as chemical, semiconductor and biology industry owing to
high flexibility and high efficiency in the production of the
low-volume, and high-value products. Process monitoring
is extremely important to ensure the safe operation of
a batch process and the quality consistency of products.
Given batch-wise variations and multiple operating phases
Manuscript received March 8, 2017; revised July 7, 2017; accepted
September 7, 2017. Manuscript received in final form September 11, 2017.
This work was supported in part by the National Natural Science Foundation
of China under Grant 61422306 and Grant 61433005, in part by the Open
Research Project of the State Key Laboratory of Industrial Control Technology
under Grant ICT170326, and in part by Postdoctoral Scientific Research
Project of Zhejiang Province under Grant 511200-X81701. Recommended by
Associate Editor P. F. F. Odgaard. (Corresponding author: Chunhui Zhao.)
S. Zhang and C. Zhao are with the State Key Laboratory of Industrial
Control Technology, College of Control Science and Engineering, Zhejiang
University, Hangzhou 310027, China (e-mail: chhzhao@zju.edu.cn).
F. Gao is with the Department of Chemical and Biomolecular Engineering,
The Hong Kong University of Science and Technology, 999077 Hong Kong.
Color versions of one or more of the figures in this paper are available
online at http://ieeexplore.ieee.org.
Digital Object Identifier 10.1109/TCST.2017.2755580
characterized by the finite duration in batch processes, batch
process monitoring is a challenging task and has attracted
increasing attention of many researchers.
Process monitoring can be divided into two distinct but
complementary steps including: 1) fault detection which aims
to identify the presence of a fault in the process and 2) fault
diagnosis which aims to isolate the faulty variables and
determine the root causes of the detected fault [1], [2]. The
fault detection approaches are, in general, classified into two
categories including the model-based method [3] and the
data-driven technique [4]. Model-based process monitoring
method requires in-depth process knowledge to characterize
the complex physical, chemical, and biological phenomena
in processes, which is hard to obtain especially for batch
processes. With the development of computers and sensors, the
abundant process data can be obtained, and thus multivariate
statistical techniques [5] such as principal component analy-
sis (PCA) and partial least squares (PLS) have been proposed
for fault detection. More recent data-driven work [6], [7] has
been reported, including real-time implementation of fault-
tolerant control systems with performance optimization and
data-driven process monitoring based on modified orthogonal
projections to latent structures. For batch process monitoring,
several fault detection methods based on the extensions of the
conventional PCA/PLS have also been reported [8]. Multiway
PCA (MPCA) and multiway PLS were proposed to handle
the 3-D data structure of batch processes. But these multi-
way methods cannot reveal the time-wise changes of process
characteristics since they take the entire batch data as a single
object. Besides, the unknown future data have to be estimated
because the entire new batch data are not available, which
may compromise the accuracy of online monitoring. In fact,
batch processes are typically conducted in a series of steps
called multiple phases with significantly different underlying
behaviors. Therefore, many phase-based methods have been
proposed to explain the local process behaviors within each
phase. Lu et al. [9] proposed a subPCA monitoring method
andthenZhaoet al. [10] proposed a soft-transition multiple
PCA method to deal with hard phase division in Lu’s work.
Based on these works, many phase-based methods have been
well developed and widely used to handle different problems
in batch process monitoring with multiphase nature [11]–[13].
Although numerous fault detection methods have been
proposed for batch process monitoring, most of the existing
methods have always inherently assumed that the modeling
batches are sufficient to cover all statistically normal batch-
to-batch variations. It is relatively easy to run trials for those
batch processes with short batch duration and inexpensive
1063-6536 © 2017 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.
See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
剩余14页未读，继续阅读
资源评论
