Reliability and performance analysis of safety-critical system using transformation of UML into state space models

https://doi.org/10.1016/j.anucene.2020.107628Get rights and content

Highlights

  • Reliability and performance analysis are the two main components of NFR analysis.

  • The framework uses state-space model to analyse system’s behavior dynamically.

  • A system that does not meet the target NFR is considered to be a failure.

  • The framework is validated with 32 critical system instances of the NPP on the RCICS .

Abstract

Non-functional requirement analysis is the most critical factor for safety-critical system construction as it reduces the risk of catastrophic loss of assets by taking measurable actions in the design phase. Reliability and performance analysis are the two main components of non-functional requirement analysis. In this paper, a reliability analysis framework is devised, which maps the Unified Modeling Language (UML) state chart model into the state-space model to analyze dynamic behavior and state transition probabilities of a safety-critical system. A system that does not meet the target reliability and performance requirements is considered to be a failure. The proposed framework is validated with 32 safety-critical system instances of the Nuclear Power Plant on the Reactor Core Isolation Cooling System module.

Introduction

Safety-Critical System (SCS) comprises of the components whose failure may lead to severe to the society in many aspects such as financial, environmental, life loss, and many more on a large scale. Nowadays, SCSs are used to serve some critical functionalities of mankind. Medical devices, power generation, aviation, railways, military equipment, etc. are just a few names in the list. Therefore, failure of the SCS (Kumar et al., 2017a) in those areas can cause a catastrophic loss in society. Thus nonfunctional requirement analysis plays a crucial role in designing those applications to reduce the risk of damage. Also, the validation of those non-functional requirements is difficult in practice. Comparatively, fewer efforts are invested in these requirements considering it as a non-viable methodology of “Fix-it-later”. Such research practices must be unacceptable to SCSs due to the risk concerns.

In nonfunctional requirement analysis, reliability and performance analysis are the two major aspects that must be ensured. In this paper, UML is considered to be the modeling language due to its power to capture all the requirements and easy understandability to the stakeholder to design the model SCSs. Specifically, the UML state chart model is a great tool to model such a scenario. The state chart model is then converted to Petri Net, and Petri Net is used to analyze through Markov chain construction with transition time using transition rate throughput. Petri Net is a mathematical modeling tool for system analysis where concurrency, choice, and iteration exist.

The framework is analyzed on different thirty-two (32) critical system instances of Nuclear Power Plant (NPP). The complete process is illustrated on the Reactor Core Isolation Cooling System (RCICS) of NPP. The paper is structured in the following scheme. Section II describes the mathematical preliminaries with the related work required for the dynamic analysis of a system. A Complete case study of RCICS is discussed in Section III. Section IV and V describe the proposed approaches for reliability and performance analysis of the SCS, respectively. Section VI describes the experimental validation of the proposed model. In section VII, we conclude the paper.

Section snippets

Mathematical preliminaries

Petri Net: Petri Net (Murata, 1989) is formally defined using 5- tuple as:PN=<P,T,F,W,m0>Where

  • P is a finite set of places {P0,P1,P2,,Pn}.

  • T is a finite set of transitions {t0,t1,t2,,tn}.

  • F is a finite set of arcs F:F(P×T)(T×P).

  • W:PXTTXPN is the arc weight mapping, where N is a nonnegative integer such that Wf>0forallfF and Wf=0forallf is not in F.

  • m0 is the initial markings. A marking of a Petri Net is a mapping. M:PN where Nis a nonnegative integer. The marking assigned to each place is

Case study

To demonstrate the proposed framework Reactor Core Isolation Cooling System (RCICS) of Nuclear Power Plant (NPP) is considered as the subject of the case study.

The proposed framework for reliability analysis with a case study illustration

In this section, a framework is proposed to carry out a reliability analysis of the SCS. Our proposed reliability analysis framework aims to conquer the limitations of existing approaches identified in (André and Benmoussa, 2014, Christine et al., 2011, Liu et al., 2013, Pettit and Gomaa, 2006, Kumar et al., 2017b, Singh and Rajput, 2016, Singh and Tripathi, 2014, Ahmad and Khan, 2013). The proposed framework is beneficial to measure all types of possible threats, which can be the root cause of

Performance analysis

A safety system must meet the high-performance requirements. To perform the performance analysis for validation of the design of SCSs of NPP, Petri Net is considered to be a good choice due to its ability to predict performance when all the system characteristics are unknown and vaguely understood. This predicted value avoids delays in system development by saving a significant amount of effort.

The average delay of a subsystem when the system in steady- state condition can compute using

Experimental validation

In this section, we estimate the rate of failure for confirmation of the proposed method experimentally by using the Brown and Lipow Input Model (Kumar et al., 2017c, Brown and Lipow, 1975). In this process, the input domain is divided into several disjoint classes. To estimate the reliability of our model, the following equation is required:Rt=i=010finiPEiWhere-

  • P(Ei): Probability specified from the operational profile. Here input course must be selected from each comparable class.

  • ni: Number

Conclusion

The approach concentrates on a technique for computing reliability and performance of the safety-critical system’s software using UML, which directly maps to Petri Nets. We obtained the reliability accuracy of our approach as 98.354511%. We have also shown the effectiveness of our approach by computing the performance of the software system. This method helps designers for early prediction of reliability during the architectural design phase itself to remove any design aspect. Therefore we

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

References (19)

There are more references available in the full text version of this article.

Cited by (9)

View all citing articles on Scopus
View full text