A review of process fault detection and diagnosis: Part II: Qualitative models and search strategies

https://doi.org/10.1016/S0098-1354(02)00161-8Get rights and content

Abstract

In this part of the paper, we review qualitative model representations and search strategies used in fault diagnostic systems. Qualitative models are usually developed based on some fundamental understanding of the physics and chemistry of the process. Various forms of qualitative models such as causal models and abstraction hierarchies are discussed. The relative advantages and disadvantages of these representations are highlighted. In terms of search strategies, we broadly classify them as topographic and symptomatic search techniques. Topographic searches perform malfunction analysis using a template of normal operation, whereas, symptomatic searches look for symptoms to direct the search to the fault location. Various forms of topographic and symptomatic search strategies are discussed.

Introduction

Diagnostic activity comprises of two important components: a priori domain knowledge and search strategy. The basic a priori knowledge that is needed for fault diagnosis is a set of failures and the relationship between the observations (symptoms) and the failures. A diagnostic system may have them explicitly (as in a table look-up), or it may be inferred from some source of domain knowledge. A priori domain knowledge may be developed from a fundamental understanding of the process using first-principles knowledge. Such knowledge is referred to as deep, causal or model-based knowledge (Milne, 1987). On the other hand, it may be gleaned from past experience with the process. This knowledge is referred to as shallow, compiled, evidential or process history-based knowledge.

The model-based a priori knowledge can be broadly classified as qualitative or quantitative. The model is usually developed based on some fundamental understanding of the physics of the process. In quantitative models this understanding is expressed in terms of mathematical functional relationships between the inputs and outputs of the system. In contrast, in qualitative models these relationships are expressed in terms of qualitative functions centered around different units in a process. The qualitative models can be developed either as qualitative causal models or abstraction hierarchies. Fig. 1 shows the taxonomy of domain knowledge based on these two broad categories. In this part of the paper we will review the various qualitative knowledge forms shown in Fig. 1.

There are fundamentally two different approaches to search in fault diagnosis (Rasmussen, 1986): topographic search and symptomatic search. Topographic searches perform malfunction analysis using a template of normal operation, whereas, symptomatic searches look for symptoms to direct the search to the fault location. Fig. 2 shows a classification of diagnostic systems based on the search methods they employ. In this paper we discuss the various search methods as shown in Fig. 2.

Section snippets

Qualitative models

The development of knowledge-based expert systems was the first attempt to capture knowledge to draw conclusions in a formal methodology. An expert system is a computer program that mimics the cognitive behavior of a human expert solving problems in a particular domain. It consists of a knowledge base, essentially a large set of if–then–else rules and an inference engine which searches through the knowledge base to derive conclusions from given facts. Also, the tree of these if–then–else

Typology of diagnostic search strategies

There are fundamentally two different approaches to search in fault diagnosis (Rasmussen, 1986): topographic search, and symptomatic search. Topographic searches perform malfunction analysis using a template of normal operation, whereas, symptomatic searches look for symptoms to direct the search to the fault location. Fig. 2 shows a classification of diagnostic systems based on the search methods they employ.

Conclusions

In this second part, of three parts, of review paper, various forms of qualitative models such as causal models and abstraction hierarchies were reviewed. Though qualitative models have a number of advantages as discussed in this paper, the major disadvantage is the generation of spurious solutions. Considerable amount of work has been done in the reduction of the number of spurious solutions while reasoning with qualitative models. In SDGs, this is done using generation of latent constraints

References (63)

  • B Kuipers

    Qualitative simulation

    Artificial Intelligence

    (1986)
  • J.E Larsson

    Diagnostic reasoning strategies for means-end models

    Automatica

    (1994)
  • M Modarres et al.

    Function-centered modeling of engineering systems using the goal tree-success tree technique and functional primitives

    Reliability Engineering and System Safety

    (1999)
  • P.R Prasad et al.

    Structuring diagnostic knowledge for large-scale process systems

    Computers and Chemical Engineering

    (1998)
  • R Reiter

    A theory of diagnosis from first principles

    Artificial Intelligence

    (1987)
  • S.H Rich et al.

    Model-based reasoning in diagnostic expert systems for chemical process plants

    Computers and Chemical Engineering

    (1987)
  • E Sacks

    Qualitative analysis of piecewise linear approximation

    Journal of Artificial Intelligence in Engineering

    (1988)
  • J Shiozaki et al.

    An improved algorithm for diagnosis of system failures in the chemical process

    Computers and Chemical Engineering

    (1985)
  • E Tarifa et al.

    Fault diagnosis, directed graphs, and fuzzy logic

    Computers and Chemical Engineering

    (1997)
  • T Umeda et al.

    A graphical approach to cause and effect analysis of chemical processing systems

    Chemical Engineering Science

    (1980)
  • R Vaidhyanathan et al.

    Digraph-based models for automated HAZOP analysis

    Reliability Engineering and Systems Safety

    (1995)
  • H Vedam et al.

    Signed digraph based multiple fault diagnosis

    Computers and Chemical Engineering

    (1997)
  • V Venkatasubramanian et al.

    An object-oriented two-tier architecture for integrating compiled and deep-level knowledge for process diagnosis

    Computers and Chemical Engineering

    (1988)
  • N.A Wilcox et al.

    Possible cause and effect graphs (PCEG) model for fault diagnosis I. Methodology

    Computers and Chemical Engineering

    (1994)
  • N.A Wilcox et al.

    Possible cause and effect graphs (PCEG) model for fault diagnosis II. Applications

    Computers and Chemical Engineering

    (1994)
  • R.G Bennetts

    Comment on the evaluation of fault trees

    IEEE Transactions on Reliability

    (1974)
  • C.C Chang et al.

    On-line fault diagnosis using the signed directed graph

    Industrial and Engineering Chemistry Research

    (1990)
  • I Chang et al.

    Model-based approach for fault diagnosis I. Principles of deep model algorithm

    Industrial and Engineering Chemistry Research

    (1994)
  • I Chang et al.

    Model-based approach for fault diagnosis II. Extension to interval systems

    Industrial and Engineering Chemistry Research

    (1995)
  • E Charniak et al.

    Introduction to artificial intelligence

    (1984)
  • F.E Finch et al.

    Narrowing diagnostic focus using functional decomposition

    American Institute of Chemical Engineers Journal

    (1987)
  • Cited by (1056)

    • Bond Graph-CNN based hybrid fault diagnosis with minimum labeled data

      2024, Engineering Applications of Artificial Intelligence
    View all citing articles on Scopus
    View full text