Risk Assessment Tools for Categorisation of Failure Modes of Marine Diesel Engine : A Comparative Study

Risk assessment is a vital element of most maintenance system, this is because safeguarding of equipment item requires maintenance strategies which usually depend on the degree of risk of the equipment item. In this paper two risk assessment tools; Risk Priority Number (RPN) based approach and Risk Matrix (RM) based approach, are presented for categorisation of risk of failure modes of marine diesel engine. The techniques are used to categorise failure modes into three risk levels; low, medium and high in turn and based on the risk level, maintenance strategy are assigned to each failure modes. Furthermore, the two techniques are compared and the result of the analysis revealed that, the extent of Risk matrix method similarity to the RPN approach depends on the benchmark for setting the risk level limit in the RPN method.


Introduction
British Standard dene maintenance as [1] the combination of all technical and administrative actions, intended to retain an item in, or restore it to a state in which it can perform a re-quired action.There are basically three types of maintenance namely; corrective, preventive and condition based maintenance.However, Marine machinery systems are composed of many equipment items and each of these equipment items possess dierent level of risk to the system.The degree of risk each equipment item possesses will determine the level of maintenance consideration, necessary for optimal system safety and reliability [2] at minimum cost.A mix of dierent maintenance strategies is therefore required for the maintenance of marine machinery system.The major challenge is how to determine which strategy is appropriate for a particular equipment item.In the use of risk as the basis for selecting appropriate maintenance strategy, the risk contribution of each equipment to the overall system is evaluated and then categorised using appropriate risk assessment techniques.The equipment items with high risk are generally tagged for Condition Based Maintenance (CBM) while equipment with low risk are labelled for Corrective Maintenance (CM).
Risk is dened as the combination of probability of failure of an equipment of a system and the resulting consequence due to the failure [2].
Three categories of risk assessment techniques are identied in the literature namely; qualitative, semi quantitative and quantitative [3].The commonly use method is the semi quantitative and one variant of it, is the risk matrix technique [3].
The risk matrix technique had been applied by dierent authors in prioritizing risk of failures modes of equipment items in diverse eld.
Nordgard and Samdal [4] used the approach to determine the risk of failure contribution of the dierent components of electricity distribution system in order to establish the appropriate mix of maintenance strategies for the maintenance management of the system.Nwaoha et al., [5] applied the technique to prioritize the various LNG carrier operations hazards.Haifang, et al., [6] utilized the approach to prioritize the risk of factors associated with the management of government project using private funds.Lazakis [7] used the approach to categories risk of failure of an equipment of a ship system into four level of criticality namely; low, moderate, signicant and high criticality.
Another variant of the risk assessment tool, is the RPN use within the framework of Failure Mode and Eect Analysis (FMEA).In this approach, risk is evaluated as the product of failure probability (O), the resulting consequences due to the failure (S) and the possibility of detecting the failure before it occurs (D).The use of this approach requires another technique such as ALARP (As Low As Reasonable Practicable) to set failure modes RPN values into dierent level of risk [8].
The RPN risk assessment methodology had been applied by dierent authors in the literature to categorise failure modes into dierent risk level.Santos et al., [8] used the RPN based approach to categorise risk of failures of dierent equipment items of an air conditioned chiller system of a vessel into acceptable, tolerable and unacceptable risk.Firstly the RPN of the various failure modes of the system were evaluated.The authors then use ALARP technique to categories the risk of failure modes into three groups; those scoring equal to or greater than 70 % of the maximum RPN value belonging to the un-acceptable risk group, those scoring between 40% and 70 % of the maximum RPN value belonging to the tolerable risk group and those scoring equal to or less than 40% of the maximum RPN value belonging to the acceptable risk group.Jamshidi et al., [9] applied Fuzzy RPN whilst considering several factors to evaluate risk contribution of various medical devices.The medical devices were then categorize into four risk levels; low priority, second low priority, high priority and very high priority.The authors proposed; corrective maintenance, preventive maintenance, condition based maintenance and preventive or condition based maintenance to the four risk levels respectively.
From the above literature survey, it is obvious, dierent authors had applied either risk matrix or RPN based methodologies individually as a tool for the categorization of risk of failure modes into three or more levels of risk and matched appropriate maintenance strategy for each risk level.However, in this paper the two risk assessment tool is utilized in the categorization of failure modes of marine diesel engine; a marine machinery system.Additionally, the two techniques are compared in an unbiased manner by applying a 10 point scale for both techniques as opposed to the use of 4 or 5 point scale in the literature for the risk matrix technique, in order to eectively determine their similarity and their eect in maintenance strategy selection.

Risk Assessment: Risk Priority Number (RPN) Approach
Risk assessment in the context of this paper consist of two components; Risk estimation and risk categorisation.Both components of the risk assessment is be evaluated with the use of RPN based approach.

Risk estimation
As earlier stated, RPN is expressed as the product of failure probability (O), the consequence resulting from the failure (S) and the possibility of detecting the failure (D).This is represented mathematically as follows: The tool is applied within the framework of FMEA for prioritising the failure modes of industrial systems which include the marine machinery systems.FMEA is a systematic technique for identifying failure modes of equipment items of a system, failure causes and eect of the failure in order to mitigate the eect of the failure.The origin of the FMEA is dated back to 1947 when it was developed by the United States Army and in the 1970s, the use of the technique was extended to the aviation and the automotive industries [10].Nowadays, the technique and its variant had become a popular tool in most industries; marine industries inclusive for evaluating risk of failure modes [11].
To evaluate RPN, values are assign to O, S and D by experts for each failure modes using a 10 point scale shown in Tab.1-3 respectively.

Risk Categorisation
To categorise risk of failure modes of the system into 3 or more risk level, experts or the decision makers can utilise a certain % of the maximum risk possible to dene the risk level limit [8].For example, Santos et al [8]  In this paper, a similar approach will be followed, to assign maintenance strategy to failure modes of marine diesel engine based on their individual risk level.The low risk failure modes equipment items will be candidates for CM while the medium and high risk failure modes equipment items will be candidates for PM and CBM respectively.

Risk Assessment: Risk matrix approach
The two components of risk assessment; risk estimation and risk categorisation in this paper's perspective can also be evaluated with a technique, which we identied as Risk matrix based approach.

Risk estimation
Risk is the product of probability of failure and the consequences of the failure [16].The low risk failure modes equipment will be candidates for CM while the medium and high risks failure modes will be candidate for PM and CM respectively.

Data collection and analysis
To illustrate the applicability of the two methods, for risk assessment of failure modes, a marine diesel engine was considered as a case

Risk estimation analysis
The value of risk of each failure mode is evaluated using Eq. 1 and the results obtained are also presented in Tab.10.
Tab. 10: Failure modes RPN.O and S as opposed to the RPN method that requires decision maker to evaluate decision criteria; D in addition to O and S. For future work, there is a need to develop a methodology that will enable industries and maintenance practitioners determined optimum index for their individual system failure scenario.

7 Tab. 5 :
Based on this denition, risk is quantied by estimating the probability of failure (O) and the consequences of the failure (S) as opposed to the use of three factors; O, S and D for estimating risk in the RPN based approach.In estimating risk, from this perspective, quantitative or qualitative risk assessment technique can be applied.In the qualitative technique, the use of a pre-determine scale, examples shown in Tab. 1 and Tab. 2, are applied in the rating of O and S. The scale is generally dened and administered by experts based on their own opinions.However, the approach is generally suitable when reliable data are not available for risk estimation and when the risk of failure of the system are mild and well known.The ratings of O and S of the dierent failure modes of the marine diesel engine by experts is the rst step in the Risk matrix based methodology.2.2.2.Risk categorisationHaving rated O and S of each failure modes of the system, with the scale in Tab. 1 and Tab. 2, the risk of failure modes are then categorised into dierent risk levels using a matrix of O and S.There are variant risk matrix various authors have applied in the literature for the categorisation of risk of failure modes into diverse risk level.Three examples, of risk matrix are presented in Tab. 5, Tab.6 and Tab.Hammed and Khan [17] risk matrix. of Advanced Engineering and Computation (JAEC) VOLUME: 2 | ISSUE: 1 | 2018 | March Tab.6: Nordgard and Samdal [4] risk matrix.5-7, a 5 point scale were used by the authors to developed O and S matrix.Tab. 5 and Tab.6, consist of three risk levels; green, yellow, red areas representing low, medium and high risk respectively while Tab.7 consist of four risk levels ; green , yellow, purple and red areas representing low, moderate, signicant and high risks respectively.From the above, dierent risk matrix have been developed and applied by dierent authors for the categorisation of risk of failure modes.Although, the risk matrix presented above utilises ve point scale, in this paper a 10 point scale is applied to develop a risk matrix based on Nordgard and Samdal [4] risk matrix and the developed matrix is shown in Tab. 8.The risk matrix is composed of three risk levels; the green, yellow and the red areas representing low, medium and high risks respectively as presented in Tab. 9. risk matrix, in categorising risk of failure modes, each failure mode O and S ratings is match with the risk matrix.For example, if a particular failure mode is assign O rating of 8 and assign S rating of 7. In the risk matrix in Tab. 8, move to O rating position 8 and check where it intercept with S rating position 7.The interception point is within the red area, meaning the failure mode risk is high.
study.The example had earlier been applied by Emovon et al. [18] to demonstrate the use of Multi-Criteria Decision Making (MCDM) enhanced FMEA in prioritising risk of failure c 2017 Journal of Advanced Engineering and Computation (JAEC) modes.The equipment items considered include among others; main bearing, cylinder head and crankshaft and 23 failure modes were identied from the equipment items.For each failure modes, three experts assigned ratings for the risk factors; O, S and D using the pre-determined scale in Tab. 1, Tab. 2 and Tab. 3. Consensus was reached among the experts in the rating of the risk factors and the agreed ratings are shown