RELIABILITY IMPROVEMENT IN BADAK LNG PLANT AMELIORATION DE LA FIABILITE DE LUSINE BADAK LNG

Bambang S. Santoso, Sr. Reliability Engineer Rahmat Safruddin, Process Engineer Ranto Manullang, Inspection Engineer PT Badak NGL, Bontang - East Kalimantan - Indonesia

ABSTRACT To keep maintain the position as the worlds leading LNG supplier, PT Badak NGL implements various programs to improve the reliability of its LNG plant located in Bontang, East Kalimantan. The effectiveness of PT Badak reliability improvement programs increased significantly after the set up of the Reliability Improvement Team (RIT) in 1996. This team identified the top ten reliability issues and continuously monitors the results of the strategy implemented to eliminate each issue. As a result, the Load Factor or uptime has improved from 90 (1996) to 92.5 % (1999), approaching the World Class level of 95 %. The ratio between the Maintenance Cost and the Current Replacement Value (CRV) has improved from 0.76 to 0.68 %, exceeding the World Class level of 1-3 %. The objective of this paper is to share PT Badak experience in implementing PTB reliability improvement programs with the LNG 13 Conference attendees in order to improve the reliability of the world LNG supply in general. RESUME Afin de maintenir sa position de leader mondial en fourniture de LNG, PT Badak NGL a mis en uvre plusieurs programmes visant amliorer la fiabilit de son usine de liqufaction de gas situe Bontang, Kalimantan Est. La mise en place dune Task-Force RIT (Reliability Improvement Team) en 1996 a nettement augment lefficacite des programmes damelioration de fiabilite.Cette quipe identifie les Dix Problemes Prioritaires en matires de fiabilit et observe continuellement les rsultats des stratgies appliques pour les liminer. En consquence le facteur de charge est pass de 90% (1996) 92.5 % (1999) en direction des meilleurs niveaux internationaux de 95%. Le ratio Cots de Maintenance / Valeur de Remplacement est pass de 0,76 % 0,68% dpassant les niveaux mondiaux de 1-3%. La prsentation a pour objectif de partager avec les autres participants de LNG13 lexprience acquise par PT Badak en ralisant les programmes prcdents, afin damliorer la fiabilit de la fourniture mondiale de GNL.

PS2-6.1

RELIABILITY IMPROVEMENT IN BADAK LNG PLANT

1. INTRODUCTION PT Badak NGL operates a LNG plant in Bontang - East Kalimantan, which has the highest LNG production capacity in the world. The company, a joint venture between the Indonesian Government (Pertamina), TOTAL Indonesie, BP, UNOCAL and Jilco, started with two process trains A and B in 1977. Currently, the total LNG production capacity of the eight process trains (A, B, C, D, E, F, G and H) is 24.1 mtpa or 53 million m3 per year. This ideal capacity is based on a condition that all process trains always operate at their current Maximum Sustainable Rate (MSR). The more reliable the plant and the feed gas supply the closer the actual capacity to this ideal capacity. To systematically improve the plant reliability, PT Badak formed the Reliability Improvement Team (RIT) in 1996 and the fully dedicated Reliability Engineering Group (REG) in 1999. Their function is to assess plant reliability, identify top ten reliability issues, monitor action program required to address each issue and evaluate the result. 2. RELIABILITY PERFORMANCE INDICATORS Until 1997, the main indicator used by PT Badak to assess the plant reliability, is the Plant Availability (PAV) which is defined as: PAV = (Actual Train Running Hours / Total Train Hours Entire Year) X 100 % When the RIT started to work, the Plant Availability (PAV) for 1988 - 1997 is as follows: Year PAV (%) '89 94.9 '90 91.8 '91 95.3 '92 93.8 '93 95.5 '94 95.3 '95 94.2 '96 94.7 '97 93.4

During this 10 years the precentages of the total LNG production loss associated with plant equipment problems are as follows : Mechanical Instrument Rotating Equipment Electrical : 21.0 : 2.9 : 1.8 : 0.6 % % % %

The PAV calculation is simple but does not count the difference in the individual train capacity, the transition and the part load production losses. Therefore, to provide a better method to quantify and categorize LNG production losses the RIT developed a Load Factor Reporting System in 1997 which uses the following definitions:

PS2-6.2

 Maximum Sustainable Rate (MSR) : The official individual Train LNG production rates in m3/hr as used in the annual LNG capacity projection MSR LNG Production : MSR X actual total hours in the evaluated month or year LNG Production Losses : Train LNG production losses in m3, calculated from the diference between the MSR and the measured actual production rate, multiplied with the duration in hours. Load Factor (LF) or Uptime : LF = (Actual LNG Prod / MSR LNG Production) * 100%

The LF is a measure of individual train Reduced LNG Production due to all causes (Plant Related and Outside Related). The LF is a direct measure of all factors that impact LNG production. Production Reliability Factor (PRF): PRF = [1 - Plant Related Losses / (Actual Production + All Losses)] * 100% The PRF is an indicator of Reduced LNG Production due to Plant Related losses only. The PRF is used to measure individual Train reliability performance relative to the annual reliability targets. Badak LNG Plant overall LF and PRF are calculated using the sums of all trains: MSR LNG production, actual LNG production, plant related losses and all losses. The overall LF and PRF statistics for 1997-1999 is displayed and discussed later in Chapter 7. There are 2 Major Categories and 13 Sub Categories of Production Loss defined and recorded in the PT Badak online Plant Operation Data System (PODS): A. Plant Related - Loss elements that are under the direct control and responsibility of PT Badak : 1. Scheduled Shutdown 2. Train Trip, Train Related ) 3. Train Trip, Utilities Related ) 4. Unscheduled Shutdown ) 5. On-line Repair ) Unscheduled Plant Related Production Loss 6. Mechanical Problem ) 7. Processing Problem ) 8. Utilities Problem ) 9. Export Derime Gas ) B. Outside Related - Loss elements that are under direct control and responsibility of Outside parties (such as, feed gas supply and LNG cargo shipping):

PS2-6.3

10. 11. 12. 13.

Inventory Control Idle Shutdown Reduced Feed Gas Other

The RIT/REG uses the PODS and the Load Factor Reporting System to identify the equipment or systems which cause the most production losses for determination of the appropriate counter measure. Figure 2.1. below shows the various production loss categories, transition losses and part load.
PLANT RELATED UNSCHEDULED S/D OUTSIDE RELATED

PRODUCTION RATE (M3/HR)

800 600
PLANT TRIP

400 200 0 0 1000

2000

3000

4000

5000

6000

IDLE S/D

PLANT PROBLEM

SCHEDULED S/D

MSR FEED GAS PROBLEM INVENTORY CONTROL

7000

8000

OPERATING HOURS

Figure 2.1 : Category of production losses

The difference between PAV and LF is shown in the calculation example on Table 2.1. The actual LF values always lower than the actual PAV because of the transition and part load losses. To improve the PRF, PT Badak performs all effort to reduce the Plant Related losses. The RIT/REG facilitates discussions between departments responsible for the identified reliability issues. The reliability issues discussed include: finding and eliminating the root causes of the Plant Related problems, PM/PdM optimization through pilot projects, reduction of scheduled shut down duration, extending time between overhaul, improving logistics practices, setting up reliability data base etc. Two of the top issues and the Pilot Projects will be discussed in Chapters 3 and 4.

PS2-6.4

Table 2.1 : Calculation example of PAV, LF and PRF Data (Example) MSR Production Rate : 700,000 m3/hr Down time (excluded idle s/d): 200 hrs Actual production : 5,850,000,000 m3 All losses : 283,000,000 m3 Plant Related Losses : 200,000,000 m3 Unscheduled: 50,000,000 m3 Calculation MSR Production (in 1 year) : 700,000 X 8760 6,132,000,000 m3 PAV = 8560 / 8760 = 97.7 % LF = 5850,000,000 / 6,132,000,000 = 95.4 % PRF = [1-(200,000,000/5,850,000,000+283,000,000)]X100%= 96.7 % =

Unscheduled : 50,000,000 / 6,132,000,000= 0.8 % of the MSR Production Scheduled : 150,000,000 m3 Scheduled : 2.4 % of the MSR Production

3. TOP RELIABILITY ISSUES 3.1. CO2 Removal Unit The problem in CO2 Removal Unit (Amine Unit) is one of the major cause of the Plant Related Losses in 1989 1999. The major problem in this plant is erosion corrosion that several times caused PT Badak to partially shut down the CO2 Removal Unit or the whole Train. In 1995, for example, PT Badak lost potential production equivalent to 243,000 m3 of LNG due problem in CO2 Removal Unit of Train E. In general, the losses are caused by lowering LNG production due to lower amine circulation rate when one bank of the cross exchanger was isolated for repair, shutdown amine and dehydration units while other units were still in operation by utilizing cross lean gas from other train, and extended downtime for equipment repair. Modifications of equipment were done by partial upgrading of piping material from carbon steel to stainless steel, especially reducers and enlargers around control valves. Some parts of carbon steel pipe on the rich amine side susceptible to erosion corrosion were also replaced with 304 stainless steel, which eliminated piping leaks. Modification of the inlet distributor to the perforated pipe type in Amine Flash Drum prevents the incoming solution from falling freely. The new inlet distributor, a standpipe with perforations located below the liquid level, prevents the solution from directly impinging the vessel wall (see Figure 3.1). The new inlet distributor design eliminated the erosion corrosion problem related to the inlet liquid

PS2-6.5

Wire Screen

Liquid Level

Splash Baffle Corrosion Areas

Before modification

After modification with perforated pipe

Figure 3.1 : Amine Flash Drum inlet distributor modification to perforated pipe. In Amine Stripper, the inlet distributor was modified from to H-type to Open Gallery-Type. This gallery-type inlet distributor reduced the vibration at the inlet distributor by readily releasing the CO2 vapor from the amine solution inside the amine stripper column. With the H-type inlet distributor, the two-phase flow caused by the release of CO2 inside the pipe creates vibration due to impingement effect induced by the shape of the inlet distributor. The shell and Tube Amine/Amine Cross Exchangers were also replaced with Plate and Frame type in Trains D and F to overcome the erosion corrosion problem experienced in the first MDEA solvent use. However, the replacement has been stopped due to improved condition after solvent replacement. In addition to equipment modification, the amine solvent was also converted from formulated MDEA to activated MDEA. Both are MDEA-based solvents but they are different in activator. The solvent conversion took place in the period of 1997-1999 following the good results of 18-month test period in Train C. After being operated with the new amine solvent for three years, the erosion corrosion of the system has been reduced significantly. This is indicated by the absence of plant unscheduled shutdown caused by amine unit problems. Other advantages in using the new solvent are 20% lower in amine circulation rate, which contributes to 20% lower steam consumption for amine regeneration and lower erosion corrosion effect and 50% lower in amine solvent losses. The iron content in solution, as an indication of corrosion, in activated MDEA was stable at the level of 1.4-1.6 ppm compared to over 140 ppm in formulated MDEA during the first year (see Figure 3.2.).

PS2-6.6

Fe Content in Solution (ppm)

160 140 120 100 80 60 40 20 0

28 -J ul98 26 -S ep -9 8 30 -M ar -9 8 29 -J an -9 8 29 -M ay -9 8 30 -N ov -9 7

Formulated MDEA

Activated MDEA

Figure 3.2 : Comparison of iron content in solution during the first one year operating with Formulated MDEA and Activated MDEA. 3.2. Main Refrigerant Compressor Impeller Problems due to cracks of the first stage impeller of the 4K-2 MCR Compressor have been experienced since March 1994. In some cases, the problem has caused significant production loss. Figure 3.3. shows an impeller crack case which resulted in a compressor failure.

Figure 3.3 : A case of impeller crack which caused a MCR Compressor failure.

01 -O ct97

PS2-6.7

Analysis of the problem root cause is very difficult due to the problem complexity and disputes between PT Badak and the compressor manufacturer. PT Badak tried to convince the manufacturer that the original 15 vanes - 3 piece impeller needs to be totally redesigned. The manufacturer tried to collect data with special data acquisition equipment to support their argument that the failure cause is not the design. In any case the acquired data did not show any evidence of operational error. Several impeller design configuration, including the latest 17 vanes - 2 piece configuration, have been tested, installed and proposed by the manufacturer as the final solution. PT Badak accept it as a temporary solution until another failure pushed the manufacturer to come with a better solution. South West Research Institute (SWRI), USA has been involved to determine the blade natural frequencies of the latest impeller configuration by conducting impact tests. Based on these data, the blade natural frequencies are plotted in a Campbell diagram. See Figure 3.4.

2,500

Available Speed Range

2,000 18 17 1,500 16

Frequency, Hz

1,000

500

0 0 1,000 2,000 3,000 4,000 5,000 6,000

Speed, RPM

Figure 3.4. Campbell diagram for a two-piece impeller. Critical frequency ranges of are presented for 16th, 17th, and 18th multiples of operating speed.

This diagram shows that the blade natural frequency will still be excited at the 17X multiple frequency of the operating speed. A valid option is to totally redesign the impeller so that the blade natural frequency is shifted sufficiently above the

PS2-6.8

influence of the operating speed. For some reasons, the manufacturer was for long time reluctant to accept this option. In the last 2 years, careful monitoring, planning and opportunity inspection conducted on the trains, have prevented unnecessary unscheduled shut down due to the 4K-2 impeller cracks. During the October 2000 warranty shut down of the new train H, the MCR compressor was inspected. Again, cracks are found in the vane-to-back weld area of the first stage impeller which has the latest 17 vanes - 2 piece configuration. With this latest finding, finally the manufacturer agrees to design a new impeller with a higher impeller natural frequency, increased about 11 % from the previous design. This increase will eliminate the resonance in the compressor operating speed range up to 4835 RPM. As short term measure, to reduce the possibility of cracks in other trains, the compressors speed is limited to a maximum of 4,620 RPM until the new design impeller is available. This speed reduction will have no impact on the trains LNG production capacity.

4. RCM AND RBI PILOT PROJECTS 4.1. RCM Pilot Project Reliability Centred Maintenance (RCM) analysis was implemented in 1996 on 15 units of 3.5 MW Cooling Water Pump. Counting of the unit trips in the subsequent years revealed an encouraging result: YEAR 1997 1998 1999 UNIT TRIP 5X 4X 2X CAUSES 3X Electrical ; 1X Instrument ; 1X Mechanical related 2X Electrical ; 1X Instrument ; 1X Operational related 2X Electrical related

However, further analysis of the trips can not find a direct correlation between the trip counts reduction and the PM/PdM tasks change resulted from the RCM analysis. The new set of tasks is about 70 % the same with the original set of PM/PdM tasks. RCM is a comprehensive but time consuming tool to evaluate an existing PM/PdM tasks or to create new PM/PdM tasks. Implementation on the 15 pump units required 1200 manhours or one full month with 6 full timers from various departments excluded a third party consultant.

PS2-6.9

Within 3 years of RCM implementation, the maintenance cost per unit decreased significantly. This is mainly because the pump motor overhaul interval is extended from 5 to 7 years. Actually the same decision can be made without following the RCM method. The decision only requires a judgement of the person who has a good knowledge of the motor overhaul process and the motor condition before and after the overhaul. The second source of cost reduction is the strategy change on the pump impeller replacement: from time based to condition based and implementation of an open bid purchasing rather than sole source. This decision can also be taken without RCM. The most benefit of the RCM is as a training tool for PT Badak plant personnel to work together at a fixed schedule to improve understanding on system approach (instead of equipment approach) and failure analysis. The RCM method is currently used by PT Badak only as a knowledge, for PM/PdM task improvement when there is a PM/PdM related equipment failure. 4.2. RBI Pilot Project Analysing inspection interval requirement of all plant equipment can be time consuming. The Risk Based Inspection (RBI) method could be used to minimise the analysis work and to focus on the high-risk items. As we know, RBI is a method to plan inspection based on an analysis of Risk of events whose probability or consequence can be affected by the inspection. The prime concern is physical damaged caused by corrosion, erosion, fatigue, etc., rather than malfunctions due to improper operation or human error. It is also pertinent to note that Risk is a function of both failure probability and the consequences of failure. In its simple form: Risk = Probability x Consequence. PT Badak has implemented this RBI method on stationary equipment of CO2 Removal Unit of a Process Train as a pilot project. This unit consists of 93 stationary equipment (pressure vessels) and 283 piping items. Tables 4.1. and 4.2 show the Risk distribution for stationary equipment and piping items resulted from the RBI assessment. The Risk Distribution for Stationary Equipment Component items indicate that 2 items (2.15%) fall within the High Consequence/High Probability area (Criticality 1) and 10 (10.75%) items fall in the Medium/High and High/Medium area (Criticality 2). This distribution reflects the spread of the majority of the stationary items in the Low Probability and Medium Consequence (Criticality 4).

PS2-6.10

Table 4.1 : Stationary Equipment Risk Distribution PROBABILITY CONSEQUENCES LOW (59.14%) (0%) 55 (59.14%) (0%) MEDIUM (30.39%) 3 (3.23%) 26 (27.16%) (0%) HIGH (9.67%) 2 (2.15%) 7 (7.52%) (0%)

HIGH (5.38%) MEDIUM (93.82%) LOW (0%)

Table 4.2 : Piping Items Risk Distribution PROBABILITY CONSEQUENCES LOW (67.84%) HIGH (0.35%) MEDIUM (96.81%) LOW (2.83%) 1 (0.35%) 183 (64.66%) 8 (2.83%) MEDIUM (24.38%) (0%) 69 (24.38%) (0%) HIGH (7.77%) (0%) 22 (7.77%) (0%)

The Risk Distributions for piping indicate that no items fall within the High Consequence/High Probability area (Criticality 1) and 22 items (7.77%) fall in the Medium/High and High/Medium are (Criticality 2). This distribution reflects the spread of the majority of piping items in the Medium Consequence and Low Probability area (Criticality 4). The interval of inspections can be determined through a scheduling tool that uses grade (confidence factor), probability (likelihood) and consequences of failure. The Inspection Grade indicates the inspection interval depending on PT Badak

PS2-6.11

confidence in the criticality rating. Table 4.3 shows the inspection intervals in months. Based on the Risk distribution of Stationary Equipment, we have two stationary equipment that falls in criticality 1. The confidence level of both equipment refer to our experience are grade 1. Therefore, the interval inspection for this equipment is 36 months. By using the same way we can determine the inspection interval for all of the equipment that has been assessed by Risk Based Inspection method. The RBI method will be continued to be used by PT Badak to determine inspection intervals of other stationary equipment of one process train in 2001. Table 4.3 : Inspection intervals (in months) Criticalitys 1 2 3 4 5 Deterioration Grade 0 24 24 36 48 60 Grade 1 36 36 48 60 72 Grade 2 N/A 96 96 144 180 Grade3 N/A N/A 240 300 360 Consequence H H/M H/M/L M/L L

High (H) Medium (M) Low (L) Negligible

5. RELIABILITY AND SAFETY In achieving a maximum result from reliability and safety improvement efforts, the RIT/REG works closely with PT Badak Process Safety Management Committee. The latest result is the integration between Project Procedure (PP Guide 0305-01) and Management of Change (MOC) Procedure through revisions of both procedures. With this improvement, all plant modification projects will be better controlled and documented. No modification, start up of new installation nor project closing is allowed prior to meeting a predetermined M.O.C. criteria. A procedure on Mechanical Integrity, another element of PSM, will be revised and improved in 2001 under the REG coordination. Equipment which is critical for safety as well for production reliability will be included in the scope of Mechanical Integrity implementation program which is one candidate of the Top Reliability Programs in 2001.

PS2-6.12

6. LIFE CYCLE COST A survey quoted by Ron Moore [1] found, while expenditures for plant and equipment occur later in the acquisition process, most of the life-cycle cost of a plant is committed at the preliminary stages of the design and acquisition process. The survey also reported that about 60-75 % of the life cycle cost is associated with maintenance and support. This means that design and installation efforts can have a dramatic effect on maintenance cost and therefore on life cycle cost. The life cycle cost concept is shown in Figure 6.1. The dotted line shows the life cycle cost curve characteristics when the maintenance cost is high relative to the initial cost.
LOW COST
100 90

CONCEPTUAL DESIGN

95 % 85 % HIGH COST CONSTRUCTION 66 % FINAL DESIGN PRELIMINARY DESIGN

LIFE CYCLE COST (%)

80 70 60 50 40 30 20 10 0

LIFE CYCLE PHASE

Figure 6.1 : Phases of Life-Cycle Cost Commitment Being concern with this life-cycle concept, PT Badak engineers have been proactively involved in every Badak Plant Expansion Projects, using their experience to provide input on maintenance and operational aspects. The engineers have been involved in the Front End Engineering & Design (FEED), Detail Design, Factory Acceptance Tests (FAT), Construction, Commissioning and Operational Acceptance of Badak LNG Plant Expansion Projects. An example result of the involvement is the installation and expansion of the on line Machinery Monitoring Systems (MMS) for critical Rotating Equipment of train E through H. By having this system installed, the severity of problems such as the 4K-2 Compressor Impeller cracks can be assessed on line, before planning a train shut down. PT Badak engineers also involved in Life Extension Project, Debottlenecking of trains A through F and DCS retrofit project.

PS2-6.13

OPERATION & MAINTENANCE

7. BENCHMARKING AND STATISTICS The RM Group Inc. conducted reliability surveys in PT Badak in 1996 and 1999. The consultant estimation for PT Badak 1996 and 1999 performance is as depicted in table 7.1. below: Table 7.1: PT Badak World Class Performance Indicators Parameter Load Factor (LF) Maintenance Cost / CRV Planned Maintenance Unplanned down time Reactive Maintenance Injury rate (OSHA recordable) Badak 1996 90 % 0.76 % 90 %+ 58 % 1.45 Badak 1999 92 % 0.68 % 90 %+ 2.6 % 50 % 1.5 World Class 95%+ 1-3 % 90 %+ < 1% < 10 % < 0.5 Typical 80-90 % 3-6 % 50-80 % 5% 40-60 % 3-4

Note: Typical is the average score of, and World Class is the highest recorded score at single facility from, all eighty assessment performed in private industry in North America. conducted by Ron D. Moore, President of The RM Group Inc. and Alan K. Pride, Senior Engineer, EMR, Inc. Table 7.2. provides a summary of the reliability survey result. Description of the used terminology and coloured rating is as follows : INTENT Stated Policy, Mission Statement or verbal recognition by the management in common objective goal for the need to undertake the activities normally associated with a particular Element of Reliability Program. TOOL The experience either of formally documented or informal procedures, practices, methodologies and skills to perform the activities associated with the stated policy of a particular Reliability Element. ANALYSIS Systems which are in place to collect, collate, analyse, and translate the data obtained or resulting from the reliability activities into useful information for management to make decisions. INTEGRATED - The effectiveness of the interaction, communication and co-ordination of the affected groups to execute the Basic Factors in order to achieve the desired reliability objectives. Table 7.2. : PT Badak Reliability Elements and Factors Representing Best Industry Practices = Best Practices Note: For each evaluated element, the = Above Average first row and the second row are = Typical respectively showing the 1996 and = Needs Significant Improvement the 1999 survey result.

PS2-6.14

ELEMENT
COMPANY RELIABILITY PHILOSOPHY PRACTICES RELIABILITY CONTINUOUS IMPROVEMENT PLAN RELIABILITY AUDITS PRACTICES ORGANIZATIONAL STRUCT. FOR RELIABILITY PRACTICES STAFF LEVELS & EXPERTISE FOR RELIABILITY PRACTISE TEAMWORK, CULTURE PRACTICES PM PRACTICES PREDICTIVE MAINTENANCE PRACTICES PROACTIVE (RCFA, ETC) PRACTICES OVERHAUL & SHUTDOWN PRACTICES INSTALLATION & COMMISSIONING PRACTICES ENGINEERING PRACTICES QUALITY ASSURANCE (QA/QC) PRACTICES OPERATIONS PRACTICES INVENTORY PRACTICES RELIABILITY DATA BASE (MMMS, EQUIP. HISTORY, ETC) TRAINING PRACTICES

INTENT TOOLS ANALYSIS FEEDBACK INTEGRATED

1996 1999

Recent benchmarking survey among LNG plants by Shell Global Solution (SGS) confirms the following results for 1999: 1. Badak LNG plant is one of the best in asset utilization (which is for all practical purpose the same as Load Factor). 2. PT Badak NGL has one of the lowest operating cost index.

PS2-6.15

Table 7.3 below contains Production Reliability statistics in 1997-1999, taken from PT Badak Load Factor Reporting System : Table 7.3 : Reliability Performance Comparison 1997, 1998 & 1999 LNG Production & Losses MSR LNG Production Plant Related Losses: Scheduled Unscheduled Plant Sub-total % Unscheduled includes: 4K-2/3 Compressors CO2 Removal Units Prod. Rel. Factor (PRF) Outside Related Losses: Load Factor (LF) Plant Availability (PAV) Std. Cargoes shipped *) 1997 100% 4.8% 2.7% 7.5% 1998 100% 6.6% 2.7% 9.3% 1999 100% 4.5% 1.0% 5.5%

0.6% 0.5% 92.6% 0.8% 92.0% 93.4% 275.9

0.8% 0.8% 90.7% 6.9% 84.1% 92.0% 290.4

0.0% 0.0% 94.6% 3.0% 92.5% 94.8% 323.4

Note: *) Standard cargo = 125.000 m3 cargo.

8. CONCLUSION Badak LNG Plant has moved forward in the period of 1997-1999, with significant progress or capability, i.e.: 1. Establishment of the Load Factor Reporting System and measurement of all losses from the ideal train production capacity. 2. Use of the Load Factor report and related information for a root cause analysis process at the executive and managerial level. E.g. solution to the CO2 removal plant and the MCR Compressor impeller problems. 3. Improvement of Production Reliability Factor (PRF) which supports continued capability for meeting customer cargo requirements.

PS2-6.16

REFERENCE CITED [1] Moore, Ron. "Making Common Sense Common Practice: models for manufacturing excellence", Gulf Publishing Company, Houston - Texas, 1999.

PS2-6.17