Department For IinJ3II..~~J International D F I )~~~Development TITLE: by: Roughness Progressions and Appropriate Maintenance Strategies for Inter-Urban Roads in Indonesia G Morosiuk, A B Sterling and S Mahmud Transport Research Laboratory Crowthorne Berkshire RG45 6AU United Kingdom PA3506199 PA3506/99 MOROSIUK, G, A STERLING and S MAHMUD (1999). Roughness progressions and appropriate maintenance strategies for inter-urban roads in Indonesia. XXI st World Road Congress, PIARC, Kuala Lumpur, Malaysia, 3 -9 October 1999. ROUGHNESS PROGRESSIONS AND APPROPRIATE MAINTENANCE STRATEGIES FOR INTER-URBAN ROADS IN INDONESIA by G Morosiuk, A B Sterling and S Mahmud 1. INTRODUCTION The accurate prediction of the rates of deterioration of roads is important in scheduling appropriate maintenance activities in road management systems. The Integrated Road Management Systems (IRMSs) project currently being undertaken by the Directorate General of Highways (Bina Marga) is attempting to integrate the various road management systems in use in Indonesia (N D Lea International Ltd, 1997). One of the objectives of the project is to standardise on the road deterioration relationships used in the IRMSs. The performance of sections of road in Indonesia have been monitored over the past ten years as part of a co-operative research programme between the Transport Research Laboratory (TRL), UK and the Institute of Road Engineering (IRE) in Indonesia. One of the outcomes of one of the TRL/IRE highway engineering projects was the derivation of appropriate calibration factors for Indonesian conditions (Morosiuk and Toole, 1998) for use with the road deterioration relationships in the World Bank's HDM-111 model (Watanatada et al, 1987). This paper reviews the roughness calibration factors derived in the TRLIIRE study and examines the economic implications of using appropriate roughness calibration factors in HDM-111 in predicting the deterioration of roads and their effect in determining maintenance strategies. The potential benefits from using appropriate maintenance strategies based on realistic predictions of roughness progressions are also detailed. HDM-111 was used to conduct the economic analysis in this study because HDM-4A is currently being developed and therefore was not available. However, the roughness progression relationship in HDM-4A is fundamentally the same as that in HDM-1l1 and therefore the roughness calibration factors derived in this study are applicable to the HDM-4A model. The research programme contributed to DFID's aims of increasing the efficiency of national and regional transport systems through better targeting of scarce resources, thus allowing the potential benefits to accrue to wider sections of the community. 2. ROUGHNESS CALIBRATION FACTORS FOR INTER-URBAN ROADS IN INDONESIA In the TRL/IRE study, twenty five sections of inter-urban road were monitored over a 10-year period. The roughness progressions observed on these sections of road were compared with those predicted by the roughness relationships based on the HDM-Il1 model (Paterson and Attoh-Okine, 1992). The predictive roughness relationships used in the comparison are reproduced below. Roughness progressions (Indonesia)1Mosistrngad amd Morosiuk, Sterling and Mahmud 1 For applications in which data or predictions of rutting, cracking and patching are available the recommended model, referred to in this study as the detailed model, is as follows: Detailed Model RIt = 0.98 eKgmmt RI,, + Kgp {0.98 eK9Ml m t[135(SNCK) 5NEt] + 0.143 RDSt + 0.0068 CRXt + 0.056 PATt } .. .(1) When a general model is required without knowledge of the surface distress, an alternative relationship for predicting roughness, which is referred to as the aggregate model in this study, is proposed as follows: Aggregate Model Rit = 1.04 eKgmm t{RI,, + Kgp) 263(1 + SNC)- 5NEt . (2) where Rit = roughness at pavement age t (in/km IRI) Rio = initial roughness (in/km IRI) t = pavement age since rehabilitation or construction (years) m = environmental coefficient NEt = cumulative esa at age t (millions esa/lane) SNCK = 1 + SNOC -0.00004(HS)(CRXt) for (HS)(CRXt) < 10,000 SNIC = modified structural number HS = thickness of bound layers (mm) CRXt = area of indexed cracking at time t (%) RDSt = standard deviation of rut depths at time t (mm) PATt = area of patching at time t (%) Kgm = calibration factor for environmental coefficient K91) = calibration factor for roughness progression The value of the environmental coefficient 'in'was set to 0.023 and the initial roughness, RI,, was set to 1.7 IRI in the analysis. Two calibration factors are used to calibrate the roughness progression models. One factor, Kgm, is associated with the environmental coefficient, m, and the other factor, K9gp, with the remaining terms. In HDM-111 the default values of Kgm and K9gp are 1.0. The average values of the two calibration factors derived from this study, for roads categorised by construction quality, terrain and traffic volumes, are given in Table 1. Table I Average roughness calibration factors by construction quality and traffic Roughness progressions (Indonesia)2M rsik st lnga d M h u Construction Traffic Detailed Aggregate Quality (MESA) Model Model Kgm Kgp Kgrm Kgp Well constructed roads with average to good Heavy > 0.75 1.3 0.9 1.3 1.0 asphalt surfacings in flat to rolling terrain and Light - < 0.75 2.4 1.2 2.6 1.2 free flowing traffic conditions Medium___ Poorly designed/constructed roads, exhibiting Heavy > 0.75 1.3 1.1 5.3 1.0 failures due to poor road widening and reinstatement prior to overlay, in flat to rolling Light - < 0.75 1.0 1.2 5.5 1.4 terrain and free flow traffic conditions Medium Well constructed roads in mountainous All All 7.0 1.5 7.0 1.5 regions with poor asphalt surfacings__________________ 2 Morosiuk, Sterling and Mahmud The figures in Table 1 indicate that the values of K9P are similar for both models and remain relatively constant for all situations. However, the values of Kg, cover a much wider range. The values of Kg. and K9 for well constructed roads in flat or rolling terrain, surfaced with well designed asphalt mixes and carrying heavy traffic, are approximately 1.0 for both the detailed and aggregate models. This indicates that both models are capable of accurately predicting roughness on these roads without the need for further calibration. When similar roads carry light to medium traffic, the value of Kgm increases to approximately 2.5 for both models, indicating that the uncalibrated models under-predict the rate of roughness progression. This suggests that traffic effects may be over-represented in the models, and/or the difference in construction quality between roads carrying heavy and light traffic may be less significant in Indonesia than in the data set from which the model was derived. On poorly designed or badly constructed roads, the development of roughness tends to be through the growth of specific defects such as cracking, rutting and patching. The detailed model was successful at using the extent of these defects to predict the rate of roughness progression. Hence Kg. in the detailed model was close to unity for both heavy and light trafficked roads. However, the value of Kg. for the aggregate model had to be increased approximately five fold to compensate for the lack of distress terms. For roads in mountainous terrain, data were only available for roads with poor asphalt surfacings. The distresses observed on these roads were either plastic deformation or early and very severe cracking. As a result of these distresses, one of the major sources of roughness was the development of transverse corrugations. The detailed model does not include this phenomenon amongst its distress variables, resulting in the model severely under-predicting the rate of roughness progression. Similarly the aggregate model also severely under-predicted the rate of roughness progression. The values of Kg, and Kgp were 7.0 and 1.5 respectively for both the detailed and aggregate models. 3. ECONOMIC CONSEQUENCES OF APPLYING APPROPRIATE CALIBRATION FACTORS The monitoring sections used in this study were located on inter-urban roads and therefore the roughness calibration factors listed in Table 1 are applicable for these types of road. The inter-urban road management system (IRMS) employed by Bina Marga uses the aggregate model to predict the progression of roughness. Hence, this investigation centred on the factors derived for the aggregate model. The HDM-11l model was used in this analysis. HDM-111 generates a number of economic indicators. The net present value (NPV), defined as the discounted benefits minus the discounted costs over the analysis period, was used as the economic indicator in this investigation. A discount rate of 15 per cent was used for a 25-year analysis period. Four notional roads were used in the analysis. A flat, straight road at sea level represented roads in flat/rolling terrain. A hilly road, at an altitude of 1500 m with rise and fall of 60 in/km and curvature of 200 0/km, typified Indonesia's mountainous roads. The roughness calibration factors derived in this study indicated that construction quality had an important effect on roughness progression. These two types of road were therefore sub-divided into 'good' and 'poor' levels of construction, and labelled as FG (flat/good), FP (flat/poor), HG (hilly/good) and HP (hilly/poor). The thickness of the bituminous surfacing of all the roads was set at 125 mm with a subgrade CBR of 5 and a structural number of 3.3. The initial roughness of the roads at the Roughness progressions (Indonesia)3Mosukselngadamd 3 Morosiuk, Sterling and Mahmud start of the analysis period was set at 2 IRI and rainfall was taken to be 200 mm/month. Five traffic levels were used, AADT's of 1000, 2500, 5000, 10,000 and 20,000, with a growth rate of five per cent per annum. These traffic levels classified by vehicle type~are given in Table 2 together with other detailed data used in this analysis. All costs are quoted in US dollars. The vehicle characteristics and optional vehicle parameters listed in were obtained from the IRMVSs study (N D Lea International Ltd, 1997). Table 2 Traffic composition and vehicle characteristics Cars Pick- ups Vehicle Type Buses Light Trucks Medium Trucks Heavy Trucks Artic Trucks 1000 400 500 25 23 25 25 2 2500 1000 1250 63 56 63 63 5 AADT 5000 2000 25~00 125 112 125 125 13 10,000 4000 5000 250 225 250 250 25 20,000 8000 10000 500 450 500 500 50 GVW .1.3 2.6 8.0 6.0 13.0 27.0 35.0 Equivalence Factor 0 0.1 0.5 0.25 2.5 7.0 9.0 Service life 10 10 10 10 12 12 12 Hours driven per year 600 1500 3000 1500 2000 2000 2000 Km driven per year 15000 30000 100000 40000 50000 50000 50000 Vehicle price 16000 16000 50000 20000 30000 40000 50000 Tyre price 25 30 130 75 130 130 200 Labour costs 1.5 1.5 1.5 1.5 1.5 1.5 1.5 Crew costs 0 1.5 1.7 1.7 1.7 1.7 1.7 Passenger time 2.5 2.5 2.5 2.5 2.5 2.5 2.5 Petrol/lDiesel 0.18 0.18 0.18 0.18 0.18 0.18 0.18 Oil 2.0 2.0 2.0 2.0 2.0 2.0 2.0 Table 3 Optional vehicle parameters Roughness progressions (Indonesia)4Mo suk Serigad a m d Vehicle Cars Pick- Buses Light Medium Heavy Artic parameter _____ ups _____Trucks Trucks Trucks Trucks Payload (tans) 0.2 0.8 3.5 4.9 7.8 17.8 15.0 Aerodynamic drag coefficient 0.45 0.46 0.65 0.7 0.85 0.85 0.63 Projected frontal area 1.8 2.72 6.3 3.25 5.2 5.2 5.75 Driving power (metric HP) 40 43 104 74 109 116 210 Braking power (metric HP) 22 38 160 131 205 250 500 Desired speed (km/h) '95 90 85 85 85 8.5 85 Energy efficiency factor 0.9 0.9 0.9 0.9 0.9 0.9 1.0 Hourly utilisation ratio 0.45 0.65 0.45 0.45 0.45 0.45 Calibrated engine speed 3500 3300 2300 2600 1800 1800 1700 Fuel adjustment factor 1.16 1.16 1.15 1.15 1.15 1.15 1.15 FRATI00 0.268 0.221 0.233 0.253 0.292 0.292 0.179 FRATIQI 0.128 0.094 0.094 0.023 Recap cost ratio 40 40 40 40 40 Tyre rubber volume (cu din) 6.85 4.35 7.6 7.3 7.3 Base number of retreads 1.5 1.5 1.5 1.5 1.5 Spare parts COSPI 23 39 1.34 7.5 3.8 3.9 13.94 Spare parts CSPQ11 6.1 6.1 6.1 25 25 25 15.65 Spare parts Q10SP 17 17 17 ___ ___ Labour hours C01LH 46.6 77.1 293.4 1180.0 1242.0 301.0 652.5 Labour hours CLHPC 0.547 0.547 0.517 10.519 0.519 0.519 0.1 1 4 Morosiuk, Sterling and Mahmud The HDM-111 model was run with the roughness calibration, factors Kg, and Kgp adjusted to the values for the aggregate model listed in Table 1. In this study no roughness calibration factors were derived for hilly roads in good condition. For these HG roads, the calibration factors for the flat roads in good condition were used in the HDM-11l analysis. The calibration factors for light/medium traffic were used for AADT's of 1000, 2500 and 5000; the factors for heavy traffic were used for AADT's of 10,000 and 20,000. Twelve alternative responsive maintenance policies were used. Each included patching of potholes and the policies required overlays of 30, 40 or 50 mm at roughness intervention levels of 4, 5, 6 or 7 IRI. The costs of these maintenance activities used in the HDM-I11 analysis are given in Table 4. The base scenario, against which each of these strategies was compared, was routine maintenance which included only patching of potholes. Table 4 Maintenance operation costs Patching 30 mm 40 mm 50 mm Routine per sq m Overlay Overlay Overlay per km per sq m per sq m per sq m per year Financial 10 5 6 7 1500 Economic 8 4 4.8 5.6 1275 The NPV for each of the twelve maintenance policies compared with the base scenario was derived for 100 km of each of the four types of road and five levels of traffic. These NPV's relate to realistic rates of roughness progression observed on inter-urban roads in Indonesia and are referred to as the 'timely' NPV's in this paper, indicating the NPV's that would be generated by HDM-111 for a range of responsive maintenance strategies that were performed, without delay, once the intervention levels of IRI were reached. The length of the national and provincial roads in Indonesia is approximately 50,000 km. No reliable data were available on the horizontal and vertical alignment of these roads. It was estimated that 70 per cent of these roads were in flat/rolling terrain and 30 per cent were in mountainous terrain. It was further estimated that there was an even split between 'good' and 'poor' roads in the flat/rolling terrain (ie. 35% in each category) and that most of the mountainous roads were in a poor condition (25% poor, 5% good). The proportions of these four categories of road were further sub-divided by the five traffic levels. The percentages of road for each traffic level are listed in Table 5. Table 5 Proportions of inter-urban roads classified by condition and traffic levels Road _ _ _ _ _ _ __ AADT _ _ _ _ _ _ _ _ Condition 1000 2500 5000 10,000 20,000 Total Flat -Good 5% 10% j10% 5% _____ 35% Flat -Poor 5% 10% 110% 5% 15% 35% Hilly -Good 1 % 1 % j 1% 1 % 1 % 5% 1Hilly -Poor 10% 5% j5% 3% j2% 25% Roughness progressions (Indonesia)5MroikStrngadahd 5 Morosiuk, Sterling and Mahmud These percentages were then used to derive the 'timely' NPV's for an average 100 km of inter-urban roads in Indonesia for each of the twelve maintenance policies. These. NPV's are presented in Table 6. The figures in Table 6 indicate that delaying maintenance to a higher roughness level reduces the NPV's. The magnitudes of the reductions in NPV's per 100 km by delaying maintenance from an intervention level of 4 IRI to 5, 6 or 7 IRI are given in Table 7 in both absolute and percentage terms. Table 6 NIPVs for timely maintenance Overlay NPV (millions US$ per 100 kin) thickness Roughness intervention level (mm) 4 IRI SIRI 61IRI 7 IRI 30 44.60 38.73 33.03 26.78 40 46.43 41.89 36.99 32.06 50 ~~147.05 43.35 38.74 34.32 Table 7 Loss in NIPV due to delayed maintenance from 4 IRI NPRV (millions US$ per 100 kin) Overlay Roughness intervention level thickness 5 IRI 6 IRI 17 IRI (mm) NP2V %NPV j NPV -~ / 30 5.87 13.2 11.57 25.9 j17.82 39.9 40 4.54 9.8 j9.44 20.3 j14.37 31.0 50 3.70 7.9 j 8.31 17.7 j12.73 27.1 The figures in Table 6 and Table 7 show that it is far more effective to provide an overlay at a roughness intervention level of 4 IRI than at 5, 6 or 7 IRI. This is true whether the overlay is 30, 40 or 50 mm thick. One of the most common maintenance strategies on inter-urban roads in Indonesia is a 30 mm overlay applied when roughness reaches 4 IRI. The figures in Table 7 indicate that there is a loss in NPV of US$5.87 million per 100 km of road in delaying this maintenance policy until roughness reaches 5 IRI. Therefore for the 50,000 km of inter-urban road there is a potential loss in benefits of approximately US$3 billion over a 25-year period. Similarly, the potential loss in benefits rises to approximately US$9 billion by delaying a 30 mm overlay until roughness reaches 7 IRI. Maintenance activities are usually limited by lack of funds. It is, therefore, important to note that it is more effective to use a 30 mm overlay at an intervention level of 4 IRI than a thicker overlay at a higher roughness level. The predicted scheduling of the maintenance activities over the 25-year analysis period have been summarised in Table 8 for the four notional roads and five traffic levels. The timing of the first overlay for a particular road and traffic level will be the same irrespective of the Roughness progressions (Indonesia)6Mroikstlngadamd 6 Morosiuk, Sterling and Mahmud thickness of the overlay. The timing of subsequent overlays however will be dependent on the overlay thickness. As the same roughness calibration factors have been used for the FG and HG roads, the maintenance schedule for these two types of road are identical. Table 8 HDM-1l1 predicted scheduling of maintenance activities Type Traffic Overlay Roughness Intervention Levels of Level Thickness 4 IRI 5 IRI 6 IRI 7 IRI Road (AADT) (mnm) Year of Years Year of Years Year of Years Year of Years 1st between 1st between 1st between 1st between ________ __________overlay overlays overlay overlays overlay overlays overlay overlays 30 11 5 13 4.5 16 3.5 17 3.5 1000 40 .7 .7 6 7 50 __ _ _ 9 9 __ _ _ 8 __ _ _ 9 30 10 5.5 13 4 14 4 16 3.5 2500 40 8 6 7 7 50 9 __ _ _ 8 9 __ _ _ 9 FG 30 9 5 12 3.25 13 3.75 15 3 & 5000 40 7.5 6 6 5 HG 50so__ 9 7 8 7 30 11 6 13 5.5 15 5 17 4 iO,000 40 8 8 8 8. 50 9 10 10 10 30 9 5 10 5 12 4.25 13 4 20,000 40 7.5 7.5 7 7 ________ ~50 8 9 9 __ _ _ 9 30 6 2.75 7 2.5 9 2 10 1.75 1000 40 4 4 3.75 3.5 50 ____ 5 5 5 5 30 6 2.5 7 2.25 9 2 10 1.75 2500 40 4 4 3.5 3.5 50 ____ 5 5 4.75 ____ 4.75 30 5 3 7 2 8 2 9 1.75 FP 5000 40 4 3.75 3.75 3.75 50 5 5 5 5 30 5 3 7 2.25 8 2 9 1.75 10,000 40 4 3.75 3.75 3.75 50 5 5 5 5 30 5 2.5 6 2.25 7 2 8 1.75 20,000 40 4 4 3.75 3.5 50 5 5 4.75 4.75 30 5 2 6 1.75 7 1.5 8 1.5 1000 40 3 3 3 3 _______ ~~50 4 4 4 __ _ _ 4 30 5 2 6 1.75 7 1.5 8 1.25 2500 40 3 3 3 3 _______ ~~50 4 4 4 __ _ _ 4 30 4 2 6 1.75 7 1.5 8 1.25 HP 5000 40 3.25 3 3 3 _______ ~~50 4 4 __ _ _ 4 __ _ 3.75 30 4 2 5 1.75 6 1.5 7 1.5 10,000 40 3 3 3 3 _______ ~~50 4 __ _ _ 4 4 __ _ _ 3.75 30 4 2 5 1.5 6 1.5 6 1.25 20,000 40 3 3 2.75 3 _____ ~~~~~~50 4 __ _ _ 3.5 __ _ _ 3.5 __ _ _ 3.75 Roughness progressions (Indonesia)7Mroukstlngad amd 7 Morosiuk, Sterling and Mahmud The figures in Table 8 indicate the consequences of delaying maintenance to a higher roughness intervention level in terms of time. The delays in time to the first overlay being scheduled by delaying maintenance from an intervention level of 4 IRI to 5, 6 or 7 IRI are summarised in Table 9. Table 9 Time delay in postponing maintenance activities from 4 lRI In reality, maintenance is frequently not performed on time. budget constraints. HDIM can be extremely useful to assist in The most common reason is budget preparation by: * estimating the years in which maintenance treatments are likely be required * calculating the cost of delaying maintenance, as an argument in support of the budget request. However, it is essential that HIDMV is calibrated for the country/environment/roads in which it is to be used. If HDM-111 is used with the roughness calibration factors Kgmn & Kgp set to the default value of 1 .0, the predicted maintenance schedule over the 25-year analysis period is as given in Table 1 0 for the various traffic and roughness intervention levels. The difference in road geometry between the 'flat' and 'hilly' roads did not affect these predicted maintenance schedules. Comparing the figures in Table 8 and Table 10 clearly shows the difference in maintenance schedules that would be estimated over a 25-year period using an appropriately calibrated HDM-111 model (Table 8) and an uncalibrated model (Table 10). For example, on a road with an AADT of 5000, the uncalibrated HDM-11l model would estimate an overlay at a roughness intervention level of 4 IRI would be required after 14 years. However, using appropriate roughness calibration factors, HDM-11I predicts that at a roughness intervention level of 4 IRI, an overlay would be required after 9 years for well constructed roads and after 4 or 5 years for poorly constructed roads. Roughness progressions (Indonesia)8M rsi ,st lnga d a m d Type Traffic Time delay (years) of Level Roughness intervention level road (AADT) 5 IRI 6 IRI 7 IRI 1000 2 5 6 FG 2500 3 4 6 & 5000 3 4 6 HG 10,000 2 4 6 20,000 1 3 4 1000 1 3 4 2500 1 3 4 FP 5000 2 3 4 10,000 2 3 4 20,000 1 2 3 1000 1 2 3 2500 1 2 3 HP 5000 2 3 4 10,000 1 2 3 20,000 1 2 2 8 Morosiuk, Sterling and Mahmud Table 10 HDM-111 predicted maintenance activities (Km & KO = 1.0) 4. SUMMARY HDM-111 roughness calibration factors, appropriate for inter-urban roads in Indonesia, were derived and used in the model to predict roughness progressions over a 25-year analysis period. HDM-11l runs were conducted for roads in flat and mountainous terrain, in good and poor condition, for a range of traffic levels using maintenance strategies of overlays of 30, 40 and 50 mm at roughness intervention levels of 4, 5, 6 and 7 IRI. The NPV's generated by HDM-111 for these maintenance strategies compared with a base scenario of routine maintenance were examined. These 'timely' NPV's indicated that it is more cost effective to provide an overlay at a roughness intervention level of 4 IRI than at 5, 6 or 7 IRI, irrespective of the overlay thicknesses that were examined. A common maintenance policy in Indonesia is a 30 mm overlay applied when roughness reaches 4 in/km IRI. This study indicated that the potential benefits that could be gained by applying this maintenance policy rather than delaying maintenance to a roughness intervention level of 5 IRI was approximately US$3 billion over a 25-year period for the 50,000 km of inter-urban road in Indonesia. The potential loss in benefits was shown to rise to approximately US$9 billion ~over 25 years by delaying this maintenance policy until roughness reached 7 IRI. The importance of using appropriate calibration factors in HDM-1ll to set maintenance schedules was illustrated in this paper. For example, the uncalibrated HDM-111 model predicted that an overlay at a roughness intervention level of 4 IRI would be required after 14 years, whereas an appropriately calibrated model predicted an overlay would be required after 9 years on well constructed roads and after 4 or 5 years on poorly constructed roads. Roughness progressions (Indonesia)9Mo sik Strnga d ah d Traffic Overlay Roughness Intervention Levels Level Thickness 4 IRI 5 IRI 6 IRI 7 IRI (AADT) mm) Year of Years Year of Years Year of Years Year o er 1st between 1St between 1St between j St betwen overlay overlays overlay overlays overlay overlays overlay overlays 30 18 - 24 - No overlay No overlay 1000 40 -- No overlay No overlay 50 - - No overlay No overlay 30 15 9 20 - 25 - No overlay 2500 40 - - No overlay 50 - - No overlay 30 14 7 17 7 20 - 23- 5000 40 9 50 9 ______ 30 11 7 14 6 16 6 18 6 10,000 40 . 9 8 9 50 10 10 30 9 6 11 5 12 5 14 4.5 20,000 40 8 7.5 8 8 50 9 9 __ _ _ 10 _ _ _ Note: the symbol indicates that no further overlays were scheduled during the 25-year period 9 Morosiuk, Sterling and Mahmud 5. ACKNOWLEDGEMENTS The work described in this paper was carried under a co-operative research programme between the Institute of Road Engineering in Indonesia and the Transport Research Laboratory of the United Kingdom, funded by the World Bank and the Department for International Development, UK. The work described forms part of TRL's research programme (Programme Director: Mr T Toole) carried out on behalf of the Department for International Development. Any views expressed are not necessarily those of the World Bank, the Department for International Development, the Directorate General of Highways (Bina Marga), or the Institute of Road Engineering. 6. REFERENCES MOROSIUK G and T TOOLE (1998). Road deterioration modelling of bituminous pavements in Indonesia -interim report. TRL Unpublished Project Report PR/ORC1131198. Transport Research Laboratory, Crowthorne, Berks, UK. N D LEA INTERNATIONAL LTD (1997). Revised Road User Cost Model Manual -Working Paper, 11'th September 1997. Integrated Road Management Systems (IRMSs). Directorate General of Highways, Jakarta, Indonesia. PATERSON W D 0 and B ATTOH-OKINE (1992). Summary Models of Paved Road Deterioration Based on HDM-111. Transportation Research Record 1344, pp 99-105, TRB, National Research Council, Washington, DC. WATANATADA T, C G HARRAL, W D 0 PATERSON, A M DHARESHWAR, A B3HANDARI and K TSUINOKAWA (1 987). The Highway Design and Maintenance Standards Model. Volume 1, Description of the HDM-111 Model. The International Bank for Reconstruction and Development, Washington, DC, USA. KEYWORDS INDONESIA, MAINTENANCE, HIGHWAY DESIGN, MATHEMATICAL MODEL, RESURFACING, FLEXIBLE PAVEMENT, ECONOMICS OF TRANSPORT, DEVELOPING COUNTRIES,1IBRD, Roughness progressions (Indonesia) 1 ooik trigadMhu 10 Morosiuk, Sterling and Mahmud