=Paper=
{{Paper
|id=Vol-2463/paper4
|storemode=property
|title=Application of a Case-Based Approach for Tasks of Industrial Safety Inspection
|pdfUrl=https://ceur-ws.org/Vol-2463/paper4.pdf
|volume=Vol-2463
|authors=Aleksandr Yu. Yurin,Aleksandr F. Berman,Olga A. Nikolaychuk, Kirill A. Kuznetsov
|dblpUrl=https://dblp.org/rec/conf/itams/YurinBNK19
}}
==Application of a Case-Based Approach for Tasks of Industrial Safety Inspection==
https://ceur-ws.org/Vol-2463/paper4.pdf
Application of a Case-Based Approach for Tasks of
Industrial Safety Inspection
Aleksandr Yu. Yurin1[0000-0001-9089-5730], Aleksandr F. Berman1[0000-0001-8339-7338],
Olga A. Nikolaychuk1[0000-0002-5186-0073], Kirill A. Kuznetsov2
1
Matrosov Institute for System Dynamics and Control Theory, Siberian Branch of the Russian
Academy of Sciences, 134, Lermontov st., Irkutsk, 664033, Russia
iskander@icc.ru
2
Irkutsk Research and Design Institute of Chemical and Petrochemical Engineering, 3,
Akademika Kurchatova st., Irkutsk, 664074, Russia
Abstract. A regular assessment of a technical state and calculation of a residual
life are important measures for the safe operation of petrochemical facilities.
These measures are parts of the industrial safety inspection (ISI) procedure. The
efficiency of this procedure can be improved with the aid of decision support
systems, which are process and reuse of accumulated information and experi-
ence on the basis of a case-based reasoning (CBR) approach. The formalization
of the main ISI tasks is made in order to apply CBR. The case models devel-
oped for each ISI task. Testing models and approach was carried out in the spe-
cial software in Irkutsk research and design institute of chemical and petroleum
engineering.
Keywords: case-based reasoning, industrial safety inspection, case models
1 Introduction
A regular assessment of a technical state and calculation of a residual life are impor-
tant measures for the safe operation of petrochemical facilities. These measures are
parts of the industrial safety inspection (ISI) procedure that regulated by different
normative documents and standards [1-6]. It is possible to improve the efficiency of
ISI with the aid of decision support systems, which are process and reuse of accumu-
lated information about earlier inspections presented in the form of cases.
In this paper we describe an example of the use of a case-based approach for deci-
sion support in the case of ISI of petrochemical facilities conducted by the Irkutsk
research and design institute of chemical and petrochemical engineering (Irkut-
skNIIHimmash). In particular, the models of cases for ISI tasks are considered.
___________________________________
Copyright © 2019 for this paper by its authors. Use permitted under Creative
Commons License Attribution 4.0 International (CC BY 4.0).
2 Background
2.1 Conceptualization and Formalization of a ISI Procedure
Most of normative documents and standards for ISI [4-6] define a composition and
ISI stages in general without specifying their content. The specific content depends on
the capabilities, experience and qualifications of the organization conducting the in-
spection [7]. In this study we use the experience of IrkutskNiiHimmash as an exam-
ple.
IrkutskNiiHimmash carries out works in the field of a technical condition assess-
ment since 1949, respectively, a large amount of information about the studies is ac-
cumulated during this period, and this information can be reused.
We define the following main ISI tasks (stages) as the results of analysis of ISI re-
ports: 1) planning works for ISI; 2) analysis of a technical documentation; 3) forming
a map of initial data; 4) development of an ISI program; 5) technical diagnostics; 6)
analysis (including interpretation) of the diagnostics results; 7) calculation of a resid-
ual resource and durability; 8) making decisions for the repair; 9) forming a conclu-
sion (report) for ISI. The information used in these tasks is sufficiently structured and
formalized, so the methods for its automated processing and reuse can be applied.
The model of the subject domain is designed as the main result of ISI conceptuali-
zation and formalization. This model includes 170 entities and relationships between
them [8]. The main information entities are the followings: "technical object", "work",
"program", "conclusion" and "decisions". These entities are detailed and the relation-
ships between them are defined. For example, elementary works are defined for the
"work" entity, these elementary works are divided into classes: documents preparation
works; technical diagnostic works; repair works and etc. Combination of elementary
works forms the investigation program for each ISI. Examples of elementary works
are: analysis of technical documentation; operational functional diagnostics; check of
readiness for technical diagnostics; visual and measuring control; thickness measure-
ment; flaw detection; study of metal properties; hydraulic tests, etc.
2.2 CBR in Petrochemistry
There are a lot of examples of the application of CBR in the field of maintenance and
diagnosing, e.g., [9-10] etc. In this work we use our early achievements in the fields
of identification of technical states of elements of petrochemical objects [11, 12], case
adaptation [13] and material selection [14].
In particular, in [11] the application of CBR for identification of technical states of
mechanical systems is considered; the main stages of the study with the reference to
the stages of the CBR cycle, metrics used, formalization of structure and properties of
mechanical systems, the structure is hierarchical composite indexes and a hierarchy of
properties of the technical objects are defined. A method for case adaptation is de-
scribed in [14]. This method is based on group decision-making methods and applied
for justification of actions for prevention of repeated failures of the petrochemical
equipment. Also we applied CBR [15] to select structural material for the design of
petrochemical constructions and parts, during this process the model of the material of
the technological object and its properties were specified.
In this work, mechanical systems and their properties are considered either as ob-
jects of a safety inspection, or as subsystems of more complex technological objects.
For this reason we can use the previously obtained results in this study. In particular,
we can consider the case as a structured representation of experience in the form of
data and knowledge prepared for its subsequent automated processing with the aid of
specialized software.
The decision-making is based on the CBR-cycle [11, 15] that includes the follow-
ing main stages: retrieve, reuse and revise. The Zhuravlev metric [16] with normaliza-
tion is used for case retrieval.
3 Models of Cases for ISI Tasks
A detailed analysis of ISI [7-8] showed the possibility to use CBR for following tasks
(stages): "development of an ISI program"; "analysis (including interpretation) of the
diagnostics results"; "making decisions for the repair" and "forming a conclusion
(report) for ISI". The different models of cases were developed for each task; these
models were based on the complete conceptual model [8]. Next, let's consider models
obtained in detail.
1) The case model for the "development of an ISI program" task. This model pro-
vides a link between the "inspection" and "program" concepts, which are included in
the descriptive part (problem) and the training part (decision) of a case, respectively
(Fig.1).
Case
+ ID: int Decision Program
+ Date: datetime
+ Type: string 1..*
Work
Operating data
Problem + Name: string Instruments and
+ Description: string + Description: string dev ices
Data of technical
Technical obj ect
diagnostics
Type of w ork
Results of control Repair information
Technical characteristics
+ Name: string
+ Value: string
Fig. 1. A fragment of the case model for task 1
The main result of case retrieval for this task is a set of similar objects and their pro-
grams selected on the basis of information about the technical characteristics and data
on the operation. On the basis of the set obtained the program for the current ISI is
formed.
Case
+ ID: int Decision Conclusion
+ Date: datetime
+ Type: string
Decisions
Operating data
Problem
+ Description: string
For technical state
Technical obj ect Data of technical
diagnostics
Results of control Repair information
Technical characteristics
+ Name: string
+ Value: string
Fig. 2. A fragment of the case model for task 2
2) The case model for the "analysis (including interpretation) of the diagnostics re-
sults" task. This model provides a link between the concepts of "inspection" and "de-
cision for technical state" concepts (Fig.2).
The main result of case retrieval for this task is a set of similar objects and results
of technical diagnosis. On the basis of the set obtained the conclusion about the
causes of the current technical state of the ISI object is formed.
3) The case model for the "making decisions for the repair" task. This model pro-
vides a link between the concepts of "inspection" and "decision for repair" concepts
(Fig.3). The main result of this task is a set of similar objects and results of repair.
4) The case model for the "making decisions for the repair" task. This model pro-
vides a link between the concepts of "inspection" and "conclusion" concepts (Fig.3).
Aggregated conclusions of similar ISIs are used as templates for the formation of
the conclusion for the current ISI. If retrieved no similar cases then selected conclu-
sion the object which belongs to similar kind of objects, e.g., container, vessel, etc.
Case
+ ID: int Decision Conclusion
+ Date: datetime
+ Type: string
Decisions
Problem Operating data
+ Description: string
For repair
Technical obj ect Data of technical
diagnostics
Results of control Repair information
Technical characteristics
+ Name: string
+ Value: string
Fig. 3. A fragment of the case model for task 3
Data of technical Operating data Repair information
diagnostics
+ Description: string
Results of control
Case
+ ID: int Decision Program Instruments and
+ Date: datetime dev ices
+ Type: string 1..*
Conclusion Work
Problem + Name: string Type of w ork
+ Description: string
Decisions
Technical obj ect
For repair
For technical state
Technical characteristics
+ Name: string
+ Value: string
Fig. 4. A fragment of the case model for task 4
4 Conclusion
One of the ways to improve the reliability and safety of technical systems in chemical,
petrochemistry and oil refining industries is to improve ISI on the basis of decision
support systems. Such systems provide interpretation of operation data, planning of
diagnostic works and forecasting of technical conditions.
In this paper we propose to apply a CBR approach for support decision-making at
the formation of the ISI program, conducting technical diagnostics, forecasting the
technical conditions; making decisions on repair and formation of the ISI conclusion.
It should be noted that the proposed approach can be used for solving another tasks,
in particular, automatic filling of sections of the ISI report and the formation of rele-
vant documents (for example, within the "analysis of a technical documentation"
task).
Algorithms and models presented were used for implementation of specialized
software [7] (Fig.5). The main effect of its application is achieved both in research to
identify laws of change in the technical conditions of the objects under consideration,
and in the organization and conduct of ISI.
Fig. 5. An example of a GUI form of software with the results of case retrieval
5 Acknowledgments
The reported study was partially supported by RFBR projects 18-07-01164, 18-08-
00560.
References
1. API 580. Recommended Practice for Risk Based Inspection,first ed. American Petroleum
Institute, Washington D.C., USA (2009).
2. API 581. Risk Based Inspection Tecnology, second ed. American Petroleum Institute,
Washington D.C., USA (2008).
3. Bertolini, M., Bevilacqua, M., Ciarapica, F.E., Giacchetta, G.: Development of risk-based
inspection and maintenance procedures for an oil refinery. J. Loss Prev. Process Ind. 22,
244–253 (2009).
4. ISO 31000:2009 Risk management principles and guidelines; 2009. 24 p. Geneva.
5. National Standard of Russian Federation 12.0.010-2009. Occupational safety standards
system. Occupational safety and health management systems. Hazard and risks identifica-
tion and estimation of risks. Moscow (Russia): Standartinform; 2011. 20 p. [in Russian].
6. Standard of the Republic of Kazakhstan 1.56-2005 (60300-3-9:1995, MOD) Risk man-
agement. System of dependability management. Risk analysis of technological systems.
Astana (Kazakhstan): National Standard of the Re-public of Kazakhstan; 2005. 63 p. [in
Russian].
7. Berman, A.F., Nikolaichuk, O.A., Yurin, A.Yu., Kuznetsov, K.A.: Support of Decision-
Making Based on a Production Approach in the Performance of an Industrial Safety Re-
view. Chemical and Petroleum Engineering 50(1-2), 730–738 (2015).
8. Yurin, A.Yu., Dorodnykh, N.O., Nikolaychuk, O.A., Berman, A.F., Pavlov, A.I.: ISI mod-
els, Mendeley Data, v1 (2019). doi: http://dx.doi.org/10.17632/f9h2t766tk.1
9. Portinale, A. L., Magro, D., Torasso, P.: Multi-modal diagnosis combining case-based and
model-based reasoning: a formal and experimental analysis. Artificial Intelligence 158(2),
109–154 (2004).
10. Liao, T.W., Zhang, Z.M., & Mount, C.R.: A case-based reasoning system for identifying
failure mechanisms. Engineering Applications of Artificial Intelligence 13, 199–213
(2000).
11. Nikolaychuk, O.A., Yurin, A.Y.: Computer-aided identification of mechanical system’s
technical state with the aid of case-based reasoning. Expert Systems with Applications 34,
635–642 (2008).
12. Berman, A.F., Nikolaychuk, O.A., Pavlov A.I., Yurin, A.Y.: Provision of Safety for Tech-
nological Systems with the Aid of Case-Based Reasoning. In Case-Based Reasoning:
Processes, Suitability and Applications (Editors: Antonia M. Leeland). NY. Nova Science
Publishers. 2010.
13. Yurin, A.Yu.: Group Decision Making Methods for Adapting Solutions Derived from
Case-Based Reasoning. Scientific and Technical Information Processing 42(5), 375–381
(2015).
14. Berman, A.F., Maltugueva, G.S., Yurin, A.Yu.: Application of case-based reasoning and
multi-criteria decision-making methods for material selection in petrochemistry. Proceed-
ings of the Institution of Mechanical Engineers, Part L: Journal of Materials: Design and
Applications 232(3), 204–212 (2018).
15. Aamodt, A., Plaza, E.: Case-Based reasoning: Foundational issues, methodological varia-
tions, and system approaches. AI Communications 7(1), 39–59 (1994).
16. Zhuravlev, I. Yu., & Gurevitch, I. B.: Pattern recognition and image recognition. In Yu. I.
Zhuravlev (Ed.), Pattern recognition, classification, forecasting: Mathematical techniques
and their application, issue 2(pp. 5–72). Moscow: Nauka (1989).