On risk-based maintenance: A comprehensive review of three approaches to track the impact of consequence modelling for predicting maintenance actions

Abstract

The integrity of the gas distribution network is crucial to guarantee the safety of human beings and the environment, while avoiding significant financial outlay. Since gas plants are progressively increasing near urban areas, a comprehensive tool to conduct maintenance and reduce the risk arising from the operations is required. To this end risk mitigation strategies have played a pivotal role during the last decades. In this paper, a comparison of three Risk-Based Maintenance (RBM) methodologies able to point out the most critical components, is presented. The first developed technique is a four stages Probabilistic Risk Assessment (PRA), characterized by a Hierarchical Bayesian Network (HBN) to perform the occurrence analysis and a Failure Modes, Effects and Criticality Analysis (FMECA) to assess the magnitude of the adverse outcomes. The HBN is adopted to overcome the limitations of traditional probability analysis approaches such as Fault Tree (FT), Event Tree (ET) or Bow-Tie (BT). To define a risk metric the total cost of failure is estimated and subsequently the Cost Risk Priority Number (CRPN) is calculated for each equipment. The second approach is a Quantitative Risk Analysis (QRA) carried out via a software named Safeti (by Den Norske Veritas -German Lloyds DNV-GL). By exploiting standard frequencies and modelling the losses of containment through Safeti, the most compelling devices are determined based on their estimated risk integral percentage. At last, Synergi Plant (another software developed by DNV-GL) is adopted for the third methodology. The software provides a Risk-Based Inspection (RBI) plan, through which the components are ranked. The proposed study can provide asset manager a concrete aid to focus maintenance efforts on priority apparatus, while assisting them in adopting the most appropriate methodology to their context. To demonstrate the applicability of the approaches and compare the obtained rankings, a Natural Gas Regulating and Measuring Station (NGRMS) is considered as case of study. The results proofed that all the proposed approaches can be implemented for practical application and the choice of the method strongly depends on the available data.

Fuzzy Risk-Based Maintenance Strategy with Safety Considerations for the Mining Industry

“…Over the last decades, maintenance has undergone a profound transformation, moving from a corrective to a proactive or predictive approach [9]. The need for this transformation stems from the fact that maintenance decisions nowadays increasingly concern complex industrial systems, and significantly affect the competitiveness and productivity of a company.…”

“…The RBM approach has been successfully used for maintenance decisions regarding mechanical and civil infrastructure, onshore and offshore structures, and cross-country pipelines [19]. In particular, it is used in sectors that are considered to be hazardous, such as, e.g., the oil and gas sector [9,[20][21][22] or the energy sector, including the maintenance of complex systems in power plants [17,[23][24][25]. An example of risk-based predictive maintenance for safety-critical systems is presented in [26].…”

“…Additionally, primary maintenance decisions are subject to uncertainty related, among other things, to uncertainty in the data, errors in the reasoning process, or the occurrence of human error [39]. This problem is also connected with selecting the best maintenance strategy for technical systems [9]. Following the identification of this issue, many methods were developed for understanding, modeling, and managing uncertainty in risk, reliability, and maintenance.…”

Rank-based risk target data analysis using digital twin on oil pipeline network based on manifold learning

“…So, the latency and traffic congestion rate will be reduced automatically. 41–43 From the MySQL database, the real-time virtual data analysis starts to proceed with the manifold learning through diffusion map tests and in order to analyse the signal strength in the frequency domain, periodogram will also be generated and stored in the array. The cloud stored data will be documented in the form of better visualization using manifold learning will java coding to get well-automated trends and pattern blocks for resemblance process and data aggregation to prevent cloud burst The patterns will be autocorrelated with the normal pattern by the cluster formation as in the real-world evaluation through the creation of Digital Twin platform with the Cloud backup to perform data mining in the predominant scope with high precision and accuracy.…”

Risk Evaluation: Brief Review and Innovation Model Based on Fuzzy Logic and MCDM

“…One of the fundamental tasks in the asset management of engineering systems (hereinafter ES) is to develop a risk calculation and prediction model. Assessment of risk level is a significant overall indicator of ES state, an indicator of remaining capabilities of ES, and guideline for risk-based maintenance methodology [1][2][3]. According to standard ISO 31000:2018 [4], risk is defined as the "effect of uncertainty on objectives".…”