• Tiada Hasil Ditemukan

Conclusions

In document .sft!f Sup<';,.,, (halaman 176-183)

7. CONCLUSION AND RECOMMENDATION

7.2 Conclusions

The proposed framework presented in this thesis is found effective in analysing gas processing systems at operation phase for improving their operational and maintenance performances. Three proposed frameworks for applying reliability;

maintainability; and availability analysis, as demonstrated through real industrial case studies, can be potentially applied by management as a strategic tool for assessing current operation and maintenance conditions, identifying weaknesses in the system and deciding on the best improvement option. Each one of the analysis can be performed separately or can be integrated into a comprehensive RAM study for overall improvement of system performances.

157

Non-repairable

Items

Continue to next page

Expert opinion elicitation

Continue to next page

I

Availability Analysis

~

Problem, objective and system definition

Continue/rom

~

previous page Data gathering

~

Model assumptions

Reliability measures

' ~

RBO model construction Maintainability

measures 1

-j_

Verified?

Yes No

T

Simulation run

j_

Validated?

No

Yes

1

Experimental design L ' .I Simulation runs

~

I

Analyze output

1

l

More runs?

Yes Yes

NoT

Make Inference and decision

~

Recommendations for Operation and Maintenance improvement

Figure 7.1: Proposed integrated framework for reliability, maintainability and availability analysis ofgas processing system at operation phase

The overv1ew of an integrated framework of reliability, maintainability and availability analysis was presented in Chapter 3 (Figure 3.1 ). A more comprehensive description of the proposed integrated framework is shown in Figures 7.1. The study should start with clear objectives before proceeding to further detailed analysis. Both reliability and maintainability (R&M) analysis follow a general flow described in Figure 3.2, which include exploratory and inferential analysis to produce estimation of R&M measures. In case of insufficient data, expert opinion elicitation method based on Figure 5.22 can be employed. The results of R&M measures estimation are then input into the availability modelling during the RBD construction step. In availability analysis, simulation technique is used to estimate system's availability and perform "what-if' scenario improvement options. The results of R&M and availability analysis can be used to assess the system and recommend appropriate actions to improve its performances.

The case study approach used in this research managed to expose actual and new problems faced by industries and in the analysis processes. Issues such as insufficient data and data with non-monotonic improvement trend have been highlighted and addressed in the study. Although the frameworks were formulated and demonstrated on the two gas processing systems, they are equally applicable to other systems in oil and gas industries.

Plant personnel involvement in the analysis processes, particularly experts was found crucial in the implementation of the proposed frameworks. The level of engagement varied across the analysis depending on the field data conditions and system complexity. Besides general tasks of providing inputs on data collection, verification and classification processes and validating model, the roles of field experts was extended to provide valuable assessment on effectiveness of maintenance improvement actions and estimation on data distribution. This level of participation has provided a good platform for engineers to incorporate their tacit knowledge on plant operation into reliability analysis systematically. Direct involvement of field personnel during the study also assisted in promoting the applications of reliability analysis and techniques in the organization under studied, which is currently still relatively low.

7.2.1 Observations from this Research

This research reveals several important points and real issues related to managing and performing reliability analysis in the industry. These points should be considered in any reliability related study, to ensure the analysis is effective and the organization gains the most benefits. The following are highlights of key findings of this research.

i) The collection, organization and verification of field data are the most critical components of reliability analysis processes since they determine the quality and usefulness of the results. These, however, are the most difficult and time consuming tasks. It is evidence from the case studies that some of the pertinent issues with field data include incomplete, missing, centralized and non-standardized data, even for an old operating plant. All of these make the analysis process tedious and more challenging. To overcome these problems, inputs from field personnel are fundamentally crucial and the use of unconventional form of data can be a good option. For example, as demonstrated in the study of AGRU, the flow rate value is used to establish the operating conditions of pumps in the absent of relevant data.

ii) Exploratory analysis plays critical roles in reliability and maintainability study of field data and should be performed in the early stage of study before more in depth statistical analyses. This analysis can provide insights on the performance of the system under studied. The plots of cumulative number of failures over cumulative operating time, and cumulative downtime numbers over cumulative downtime duration, for example, are found very useful in gauging current system operational patterns, providing clues on outlook of system performance and identifying suitable mathematical model for predicting future trends.

iii) Maintainability analysis is found to be as critical as reliability analysis for system at operation phase since it can be a dominant factor influencing the availability performance, as demonstrated in the gas compression train system case study. The findings from maintainability analysis will reveal the overall effectiveness of maintenance system and improvement actions. Important

161

attributes from the findings such as lesson learned, best practices and effective mitigation actions should be well shared not only with others at operation but also be feedback to the design team responsible for the development of similar systems in future. Proven improvement program addressing logistics issues such as vendor support and spare parts availability should be well established before the system commences operation.

iv) The assumption of constant failure and repair rate (random event) has to be tested first by means of statistical analysis, before it can be applied, even though it is generally acceptable for failure rate at system level due to the effects of various subsystems and components. Similarly an liD test should be performed on the field data before they can be analyzed using life data analysis (LDA) approach.

v) Reliability data analysis is not a well-known technique among engineers and management even though they are aware of the importance of having reliable operation. The analysis is usually done on ad-hoc basis and generally suffers from unstructured and unsystematic approach. Issues such as inadequacy, poor reliability and traceability of the existing database usually cause the studies to take longer time to complete and the results to be subjected to greater uncertainty. Other concerns include lack of skills and competency in the techniques and prevailing scepticism towards statistical-based analysis results amongst management. Nevertheless, the tendency in considering reliability analysis as an important improvement tool is rapidly apparent in the organization under studied based on their good support and participation in the case studies.

7 .2.2 Contributions of the Research

This research aims to fill in the gap found in the literature on the applications of RAM analysis in the industries, particularly in oil and gas sector. Main contribution of the research is the proposed framework for implementing reliability, maintainability and availability analysis effectively to improve gas processingsystem performances.

This research further enhances the knowledge thus far in the RAM analysis of plant system at operation phase in actual industrial applications in the following areas:

1. It provides generic frameworks on how to perform and apply reliability, maintainability and availability analysis individually (Figures 3.1, 3.2 and 3.4) and collectively (Figure 7.1) as a strategic tool for plant management to evaluate existing performances, identify critical factors and overcome operational challenges.

2. It presents practical approach on how to tap, engage and exploit plant personnel and field expert's knowledge in the analysis processes (Figure 5.22).

The proposed expert elicitation method to quantitatively estimate probable distribution of maintenance data can also be applied to other situations where the data are scarce and limited.

3. It addresses the issue of predicting system performance having non-monotonic trend as a result of maintenance and operation improvement, as in the case of corrective maintenance downtime of gas compression train system.

Approaches based on linear regression and expert censoring techniques have been proposed in that situation (Figure 5.19).

4. It presents a framework (Figure 5.19) for conducting maintainability analysis at operation phase thatenhances the maintainability requirement model described by Blanchard et a/. (1995) in Figure 2.5, by looking at ways of sharing lessons learned and providing more effective feedback on operation performance to both the design team and other plant personnel working on similar system.

5. It outlines a framework for integrating availability modelling and simulation techniques to assess various operational situations and estimate availability gained, upon which can be used to assist management in the strategic decision making process (Figure 3.4).

In document .sft!f Sup<';,.,, (halaman 176-183)