6  A COST AND QUALITY OPTIMISED MODEL FOR PAVED ROADS

 

6.1 NAMIBIA'S BUDGETARY MAINTENANCE PREDICAMENT

 

Namibia's Department of Transport is in a predicament in so far that the maintenance expenditure has accelerated faster than the ever rising maintenance demands of an increasing roads system could reasonably expect. The reasons for this disturbing situation in Namibia are the same as in many other countries [49]: inflation, recession, low productivity, constant increase in traffic and axle loads to unforeseen levels and environmental factors like the specific Namibian climate. Additional aggravating circumstances in Namibia were caused by a situation of colonial power structures until very recently with the resulting unbalanced economical situation. All these factors have contributed to the situation where the budgetary maintenance situation is simply got out of hand.

In order to overcome this problem the importance of cost-efficient road maintenance has to be emphasised. With adequate road maintenance the road pavement life can be maximised, and the vehicle operating costs can be minimised. Independent Namibia has to learn from negative lessons in other African countries where road maintenance has been in many cases grossly neglected or given a low priority with the result of continuously deteriorating road networks. The basis for optimised road maintenance for paved roads will be a road pavement management system (PMS) which must contain the following data and decision-influencing facts [50]:

- Establishing priorities for maintenance, rehabilitation and reconstruction in selected areas, based on criteria set by the management of the Department of Transport;
- Obtaining feedback relative to pavement performance for input into pavement design, construction and maintenance activities;
- Applying life-cycle cost analysis when reviewing alternatives;
- Considering major design parameters like foundation strength, number of projected E80 axle loads, material specifications, climate and others.

More than one third of the system of approx. 4.500 km of paved roads has pavements which are more than 20 years of age, and the average age of the most current seals on these paved roads is 9 years. For the next 5 years the length of paved roads which are older than 20 years will increase with approx. 277 km/year. It has been proved over 30 years of maintenance experience on Namibia's paved road network that the optimum maintenance sequence will be an 11, 9, 7, 5 years reseal cycle. After completion of this cycle, after approx. 30 years, rehabilitation or even rebuilding of the paved road can be required. To keep this reseal sequence it would be required to reseal 600 km paved roads per year while between 1980 and 1986 only 190 km per year were in fact resealed. The mean cost of a reseal is in the order of US $ 16.000 per km for a 13 mm chips surfacing which represents an average value between a cheap fogspray and a full-scale premix. For an ideal maintenance tempo of 600 km/year US $ 8.000.000 will have to be budgeted for. But besides these more regular reseal requirements a road has to be rebuilt after some time. Under best circumstances it can be assumed that a road will need a complete rebuilding upgrade after 30 to 40 years of service life. With this re-building-cycle, for a current network of 4.500 km of paved roads and a current mean price of US $ 52.000 per km for a re-built road, a further US $ 4.700.000 per year are required. This total estimate of US $ 13.000.000 for paved roads maintenance is not available at the current date.

With the to-date used conservative and non-objective methods it was not possible to establish consistent maintenance priorities. Methods to monitor the conditions of the pavements in an orderly and rational manner didn't exist to date due to a lack of adequate methods and trained and motivated personnel.

 

6.2 VISUAL-SENSITIVE METHODS TO SOLVE THE PROBLEM

 

To achieve the objective of the establishment of consistent and sound maintenance priorities the Namibian Department of Transport decided during 1983 to implement a computerised " Pavement Management System" ( PMS) which has been developed by the " Division of Roads and Transport Technology" of the CSIR. The basis for such a system is to perform an inventory of the entire pavement network and to evaluate its condition. This data-collecting visual-sensitive operation is a tedious task but with a well designed rating procedure and properly trained and motivated data collecting personnel, it has been proved that this system, with some shortcomings, runs smoothly under Namibian conditions.

Although no universal system exists to evaluate and assess a road pavement quality system, the data collecting evaluations has to include:

- All pavement structure data
- Traffic counts
- Construction and maintenance history
- Drainage structures data

To achieve these objectives the Department of Transport developed a visual inspection method for paved roads including a point-allotment and seriousness rating index system, supported by a " Linear Displacement Integrator" (LDI) (see chapter 7 for a description) to measure the driving quality on pavements. The Computer Section of the Department of Transport developed, with the assistance of the 'DRTT', a reseal algorithm for the departmental computer system, a HP 1000 F mini-computer, to establish reseal priorities for paved roads. Computer models play a major role in estimating the cost tradeoffs among the various maintenance and rehabilitation strategies with consideration of the transportation costs.

To establish reseal-priorities the following basic activities are required, based on visual-sensitive methods:

1. Objective visual inspections to determine defects and other pavement properties;
2. The allotment of seriousness ratings to the defects;
3. A decision procedure to establish under which circumstances maintenance, resealing and rehabilitation will be justified. Furthermore it will be required to determine a procedure to compare the defects and other pavement properties of all pavements in Namibia and consequently develop a list of reseal and rehabilitation priorities.

With these basic requirements as starting point the first Namibian 'PMS' was developed during 1983 consisting of a procedure for visual surveys, the allotment of seriousness ratings and the establishment of a computerised algorithm to determine reseal priorities.

 

6.2.1 VISUAL SURVEYS

 

The following procedure has been developed and adapted by the Department of Transport:

1. Use a ' PMS' with test sections of 50 m each for every 5 km of paved road, but at least 3 for any road.
2. Three different types of cracks, rutting, percentage patching spots, the measure of loss of aggregate and the bleeding of the bituminous surfacing will be judged.
3. A panel of experienced road technicians has to do the survey.

Types of cracks are an important parameter for the description of road characteristics. Surfacing cracks and their properties are topics of research in many countries. It is doubtful whether one should be so audacious to specify widths of cracks. In theory this seems to be very worthwhile but it must be questioned whether this is really a realistic concept. However, it must be remembered that visual inspections can be done on a network level, or on a project level with considerable more detail. The procedure for the Namibian 'PMS' visual inspection was based on the needs for a network level. For this purpose the visual inspection must have enough detail to identify possible problem areas. It must also be ensured that the sample section is fairly representative of the relevant road piece.

The seriousness rating will be determined by the grade of cracking, the loss of surfacing aggregate and the grade of bleeding. This rating will be further influenced by the percentage of patching of a specific test section and the depth of rutting. Three different types of cracks will be investigated: longitudinal cracks, cross cracking and crocodile cracking as well as combinations of these three crack types. The seriousness rating due to cracking is a function of the percentage length of the different cracks in relationship with the total length of a test section (length of test section with two lanes is 50 m). The crack width, whether it is >1 mm or <1 mm, will also influence the seriousness rating. The seriousness rating has a range from 0 to 3 (0 = no problem, 3 = unacceptable). In the frequent case of combinations of different crack types, each crack type will be differently evaluated with the allotted seriousness rating and then summarised in table 28.

The severity indices for each mode of distress are related to the urgency of maintenance requirements which result from the presence or absence of these distress modes. The general definition which describe these maintenance requirements are as follows [86]:

Rating 0 Virtually no distress is evident and there is no requirement for maintenance or even the need for monitoring the pavement condition in the following year.

Rating 1 Notable distress that should be monitored but not yet be a requirement for maintenance in the following year.

Rating 2 Distress which significantly affects the porosity of the surface and/or indicates a reduction in the structural capacity of the pavement, and/or seriously affects the level of safety afforded by the surfacing. Such a condition would be a probable requirement for maintenance within the present or following budget year.

Rating 3 Distress which offers negligible protection against the ingress of water, or that has led to a disintegration of the surfacing, or that is in any way damaging the pavement structure, and/or indicates a hazardous condition for traffic. Such a condition would require urgent maintenance.

 

TABLE 28  SERIOUSNESS RATING FOR CRACKING

 

|==================================================================|
|
            |         SERIOUSNESS FOR CRACKING                                    |
|-------------|---------------------------------------------------------------------|
| CRACK WIDTHS| SERIOUSNESS RATINGS                                                 |
|             | PERCENTAGE OCCURRENCE: RATIO: CRACK LENGTH/TOTAL Length             |
|             |           20%|           40%|         60%|         80%|         100%|
|-------------|--------------|--------------|------------|------------|-------------|
|             |         LONGITUDINAL CRACKS (CRACK LENGTH/TOTAL LENGTH)             |
|-------------|--------------|--------------|------------|------------|-------------|
| > 1 mm      |      1       |       2      |      2     |      3     |       3     |
| < 1 mm      |      0       |       1      |      1     |      2     |       2     |
|-------------|--------------|--------------|------------|------------|-------------|
|             |        CROCODILE AND OTHER CRACKS (CRACK LENGTH/TOTAL Length)       |
|             |      2%      |       4%     |      6%    |      8%    |      10%    |
|-------------|--------------|--------------|------------|------------|-------------|
| > 1 mm      |      1       |       2      |      2     |      3     |       3     |
| < 1 mm      |      0       |       1      |      1     |      2     |       2     |
|-------------|--------------|--------------|------------|------------|-------------|
|             |         CROSS CRACKS                                                |
|             |         NUMBER OF CRACKS PER TEST SECTION (50 m)                    |
|             |      20      |      40      |     60     |     80     |     100     |
|-------------|--------------|--------------|------------|------------|-------------|
| > 1 mm      |      1       |       2      |      2     |      3     |       3     |
| < 1 mm      |      0       |       1      |      1     |       2    |        2    |
|===================================================================================|
NOTA: Department of Transport: Materials Section, 1986

The seriousness rating due to loss of surfacing aggregate and to bleeding will be evaluated according to table 29:

 

TABLE 29  SERIOUSNESS RATING: SURFACING AGGREGATE AND BLEEDING

 

|==================================================================|
|             | LOSS OF SURFACING AGGREGATE           |  BLEEDING  |
|-------------|---------------------------------------|------------|
| SERIOUSNESS | CHIP SEAL | SAND OR SLURRY |  PREMIX  | applicable |
| RATING      |judge chip |   SEAL         |   WARM   | on weakest |
|             |   only    |                |          | 50 m lane  |
|-------------|-----------|----------------|----------|------------|
|             |No or small|No or small loss|  No or   |  No or few |
|      0      |  loss of  |   of sand or   |small loss| spots where|
|             | aggregate |   slurry seal  |  of fine | bitumen is |
|             |           |                |aggregate |equal to ag.|
|-------------|-----------|----------------|----------|------------|
|             | less than | less than 10%  |<20% loss |<20% length |
|      1      |20% loss of| sand or slurry |of fine ag|of lane is  |
|             | aggregate | seal loss      | and 0% of|bit equal to|
|             |           |                |coarse ag.|the chips   |
|-------------|-----------|----------------|----------|------------|
|             |   20-50%  |  10-20% loss   |20-50% ag.|   20-75%   |
|      2      |  loss of  |                | and <20% |length: lane|
|             | aggregate |                |coarse ag.|is bit equal|
|             |           |                |   loss   |to the chips|
|-------------|-----------|----------------|----------|------------|
|             |     >50%  |    >20% loss   | >50% fine|    >75%    |
|      3      |  loss of  |                |ag.and>20%|length:lane |
|             | aggregate |                |coarse ag.|is bit equal|
|             |           |                |   loss   |to the chips|
|==================================================================|
NOTA: Department of Transport: Namibia: Directorate Planning:
Materials Section, July 1986

 

6.3  THE FIRST NAMIBIAN RESEAL ALGORITHM

 

The Namibian Department of Transport has developed a computerised algorithm (Encyclopedia Britannica: An algorithm is a set of steps designed to achieve a complex mathematical operation, each step carrying the operation forward by one small increment, and with a built-in repetition of a step until certain conditions are reached) to determine the reseal priorities. The input data into this algorithm are the different types of crackings, the loss of surfacing aggregate and the grade of bleeding with their allocated seriousness ratings. Normally only seriousness ratings of between 2 and 3 will be considered and depth of rutting of more than 15 mm. Another important input into this algorithm is the ' AADT' (Average Annual Daily Traffic). The reseal algorithm results in different actions like for instance: do nothing, reseal within 2 years or reseal immediately.

 

6.4  REVISED PROPOSALS FOR A RESEAL ALGORITHM

 

The first method to determine reseal priorities has proved to possess different inherent weaknesses which should be overcome using a revised method. These weaknesses are inter alia the following:

1. Weaknesses in the process of visual surveys, for instance the determination of loss of surfacing aggregate or the determination of aggregate loss in the case of slurry seals.

2. It is not possible to establish "priorities of priorities". In the case of longitudinal cracks everything > 50% is very serious. It is further experienced that the original 3-point system could not differentiate between "bad, but still tolerable", and "unacceptable, immediate action required".

3. The influence of Namibian climatic conditions is not considered, as the condition of bituminous binder and the permeability of the pavement.

4. The texture of the existing seal is not considered.

5. It has been proved that test sections (50 m) for every 5 km are positioned too far apart.

6. It has been also proved that the method to establish the urgency of a reseal is not good enough because the reseal priorities have in many cases been proved not to be realistic. The data obtained from visual surveys should be more realistic and an alternative method to allot seriousness ratings to defects should be found.

The revised procedure to determine reseal priorities is proposed is dealt with in the next sections.

 

6.4.1  REVISED VISUAL SURVEYS AND SERIOUSNESS RATINGS

 

More defects and properties have to be judged. Furthermore it is proposed to space test sections only 2,5 km apart in place of the originally envisaged 5 km to obtain a more realistic picture.

It is also proposed to use a five point seriousness rating scheme in place of the original three point scheme except for the judgement of the texture of the pavement and the condition of the bituminous binder.

The following broad principles in the judgements of defects will have to be followed:

 

FIGURE 5  JUDGEMENT OF PAVEMENT DEFECTS

 

Figure5.gif (92633 bytes)

Rating of Defect:

(1.): Insignificant consequences, difficult to differentiate and
      cognisable
(2.): Easy to differentiate and cognisable but with minimal or
      small consequences
(3.): Noticeable consequences, normally still acceptable
(4.): Important consequences, normally not acceptable
(5.): Serious consequences, absolutely unacceptable

 

TABLE 30  SERIOUSNESS RATING FOR CRACKING: REVISED METHOD

 

|==================================================================|
|
            |               SERIOUSNESS FOR CRACKING                              |
|-------------|---------------------------------------------------------------------|
| CRACK WIDTHS| SERIOUSNESS RATINGS                                                 |
|             | PERCENTAGE OCCURRENCE: RATIO: CRACK LENGTH/TOTAL Length             |
|             |          20%|          40%|          60%|           80%|        100%|
|-------------|-------------|-------------|-------------|--------------|------------|
|             | LONGITUDINAL CRACKS (CRACK LENGTH/TOTAL LENGTH)                     |
|-------------|-------------|-------------|-------------|--------------|------------|
| > 1 mm      |       1     |       2     |       3     |        4     |      5     |
| < 1 mm      |       1     |       1     |       2     |        3     |      4     |
|-------------|-------------|-------------|-------------|--------------|------------|
|             | CROCODILE AND OTHER CRACKS (CRACK LENGTH/TOTAL Length)              |
|             |       2%    |       4%    |       6%    |        8%    |     10%    |
|-------------|-------------|-------------|-------------|--------------|------------|
| > 1 mm      |       1     |       2     |       3     |        4     |      5     |
| < 1 mm      |       1     |       1     |       2     |        3     |      4     |
|-------------|-------------|-------------|-------------|--------------|------------|
|             | CROSS CRACKS                                                        |
|             | NUMBER OF CRACKS PER TEST SECTION (50 m)                            |
|             |           20|           40|           60|            80|         100|
|-------------|-------------|-------------|-------------|--------------|------------|
| > 1 mm      |       1     |       2     |       3     |        4     |       5    |
| < 1 mm      |       1     |       1     |       2     |        3     |       4    |
|===================================================================================|
NOTA: Department of Transport: Namibia: Directorate Roads: Materials Section, March 1990

As in the case of the original method three different types of cracks will be investigated: longitudinal cracks, cross cracking and crocodile cracking as well as combinations of these three crack types. The seriousness rating due to cracking is a function of the percentage length of the different cracks in relationship with the total length of a test section (length of test section with two lanes is 50 m) respective the number of cracks. The crack width, whether it is >1 mm or <1 mm, will also influence the seriousness rating. Again it has to be emphasised whether it is worthwhile to go so deep into detail. Scepticism is appropriate. The seriousness rating has a range from 1 to 5 and is shown in table 30.

 

TABLE 31  SERIOUSNESS RATING: SURFACING AGGREGATE AND BLEEDING

 

|==================================================================|
|
            |              LOSS OF SURFACING AGGREGATE            |  BLEEDING     |
|             |        applicable for the weakest area of at        |  applicable   |
|             |        least 10 mē, at least 1 m wide on the        |  on weakest   |
|             |                     50 m test section               |   50 m lane   |
|-------------|-----------------------------------------------------|---------------|
| SERIOUSNESS | ANY TYPE OF BITUMINOUS SEAL  |         PREMIX       |               |
| RATING      | excluded premix              |         WARM         |               |
|             |                              |        judge loss    |               |
|             |                              |        fine ag.mm    |               |
|-------------|------------------------------|----------------------|---------------|
|             | 0 to 10 equivalent holes     |        To 1 mm       |    0-20% of   |
|      1      | in the 10 mē of the          |        loss of       |    length of  |
|             | 50 m test section            |         fine         |   lane is bit |
|             |                              |        aggregate     |  equal to ag. |
|-------------|------------------------------|----------------------|---------------|
|             | 10 to 20 equivalent holes    |       To 2 mm        |   20-40% of   |
|     2       | in the 10 mē of the          |        loss of       |   length of   |
|             | 50 m test section            |         fine         |   lane is bit |
|             |                              |        aggregate     |  equal to ag. |
|-------------|------------------------------|----------------------|---------------|
|             | 20 to 30 equivalent holes    |       To 3 mm        |   40-60% of   |
|     3       | in the 10 mē of the          |        loss of       |   length of   |
|             | 50 m test section            |         fine         |   lane is bit |
|             |                              |        aggregate     |  equal to ag. |
|-------------|------------------------------|----------------------|---------------|
|             | 30 to 50 equivalent holes    |       To 4 mm        |   60-80% of   |
|     4       | in the 10 mē of the          |        loss of       |   length of   |
|             | 50 m test section            |         fine         |   lane is bit |
|             |                              |        aggregate     |  equal to ag. |
|-------------|------------------------------|----------------------|---------------|
|             | 50 and more equivalent       |         > 4 mm       |    > 80% of   |
|     5       | holes in the 10 mē of        |         loss of      |    length of  |
|             | the 50 m test section        |           fine       |   lane is bit |
|             |                              |         aggregate    |  equal to ag. |
|===================================================================================|
NOTA: Department of Transport: Namibia: Directorate Roads:
Materials Section, March 1990

An " equivalent hole" has a diameter of approx. 10 mm. Bleeding is still included to serve in the case of "serviceability limit state" problems as guideline, but is not used as structural parameter in the priority rating process for an appropriate reseal any more.

The revised seriousness rating due to loss of surfacing aggregate and to bleeding is evaluated according to table 31. Experience gained during the first few visual assessments led to the method of rating " stone loss" or " ravelling" by the number of " equivalent holes" that occur in the layer of multiple seals. Stone loss is difficult to judge, but it is experienced that the stripping problem is only serious if physical holes do appear. If, for instance, chips are lost from a second or third bituminous seal (reseals) but water cannot penetrate the underlying pavement layer due to an impermeable seal underneath the reseal, it cannot be rated as serious. For this reason it was decided to be relatively conservative and to rate "stone loss" in terms of "equivalent holes".

No formal method exists as yet in Namibia to evaluate " permeability". Many tests were made by using a falling head permeameter in trying to set standards. However, numerous problems such as lateral movement of water in the surfacing layer, the seating of the apparatus over cracks etc. led to the conclusion that it is not feasible to physically measure permeability when collecting data. At this point of time the attribute "permeability" is rated subjectively by an experienced team by using water containing a washing detergent to break surface tension. It is, however, possible that a high rating of permeability can be attained without the occurrence of extensive cracking due to a porous surfacing layer.

 

6.4.2 PROPOSED NEW RESEAL ALGORITHM

 

The main criterion to establish the appropriateness or non-appropriateness of a seal is to determine whether it fulfils its main function to protect the basecourse or the upper-most bearing course of a road structure from traffic loads and penetrating water. The following algorithm will be used to establish whether a reseal is appropriate or not:

 

FIGURE 6   APPROPRIATENESS OGF A RESEAL

 

Figure6.gif (5085 bytes)

It is absolutely essential to use fresh data for the establishment of reseal priorities. The personnel which is used to execute the pavement survey must be motivated to fulfill this task as reliably and objectively as possible. The visual survey must be always as consistent and objective as possible. An objective and reliable method to establish the permeability and the dry-condition of the bituminous binder must still be found and is not available to date.

The following revised Reseal Algorithm is proposed to determine whether a reseal will be required or not:

 

FIGURE 7  REVISED RESEAL ALGORITHM

 

Figure7.gif (16025 bytes)

The basis for any "Pavement Management System" in order to achieve optimal maintenance results is the knowledge about the structural condition and the structural layerworks of any paved road network. The resulting reseal algorithm which has been discussed above depends only on mechanistic deterioration models, on traffic counts and climatic conditions but not on vehicle operating costs which were dealt with in the previous chapter. It will be shown that the Namibian ' PMS' cannot be based on 'VOC', because 'VOC' cannot justify any rehabilitation or re-building based on 'VOC' due to the extreme low traffic loads on Namibian roads. A survey of the most important structural properties of the Namibian paved roads system as established according to the first ' PMS'-System is available on the departmental computer system.

 

6.5  STRUCTURAL ASPECTS OF PAVEMENTS

 

6.5.1  STRUCTURE OF PAVEMENTS OF PAVED ROADS

 

The structure of the Namibian paved roads has been collected in a databank 'NAMROAD' which has been compiled by the author for a 'HP VECTRA ES' PC-computer. It gives for each section of the different road classes (trunk, main and district roads) a short description about the location, the length of the section, the specific surface and shoulder widths, the shoulder material type, the average driving quality in terms of the "International Roughness Index" ( IRI) as established by the ' LDI' instrument in terms of ' PSI' and recalculated to 'IRI', the minimum and maximum ruttings in mm (measured by a 2 m long straight edge), the year of opening and the number and years of major reseal operations since the opening of each concerned section.

The structure-description of each section of the paved roads system contains the kilometre stations, the layer type, the layer thickness, the material class and the description code of the different material classes. For each major change in the structure of layer types or thicknesses or for a change in the material class or both, a new subsection will be started.

The pavement structure of all sections of paved roads in Namibia must be brought into association with the traffic loads on these sections. For each major change of traffic patterns a new subsection will be started. The traffic loads will be described by the absolute numbers of the "Annual Average Daily Traffic" ( AADT), the number of heavy vehicles as well as the percentage of heavy vehicles, the number of equivalent E80 axles per lane of the paved road and the E80 factor.

 

6.6  RESULTS OF THE PERFORMANCE MODEL STRUCTURE

 

The compressed pavement structure data as established in the computerised databank 'NAMROAD' has had one objective: to develop a performance model for the different structural pavement types to simulate the behaviour of these pavements under Namibian traffic conditions. Such a prognosis model is important due to the fact that the Namibian paved roads are quite unique, even in Africa, as far as very low traffic numbers, climatic conditions and unique natural road building materials are concerned. It must, however, be considered that traffic numbers are low in terms of ' AADT', but in many cases not in terms of E80 axles. One factor is especially beneficial for the performance of Namibian paved roads in that they are designed for a soaked CBR and not an optimum moisture CBR value. For many basecourses, for instance, a soaked CBR of 80 is specified but the basecourse is, due to the specific very dry climatic conditions in Namibia, never soaked in its life time and retains a real CBR value in excess of 200. This is an additional safety factor in the strength behaviour of pavement structures, but it will have to be considered to use a CBR value under optimum moisture instead, with the resulting relaxed design specifications and cheaper materials.

The main criterion for the behaviour of Namibian paved roads is the IRI value, measured by the 'LDI' roughness measuring instrument, which serves as standard in judging the riding quality of a paved road in Namibia and is recalculated from the original ' PSI' (Present Service Ability Index). This conversion to IRI must be seen with suspect but due to a lack of better research results the above equation has to be used preliminary with reservations as outlined previously. The inclusion of the rutting value in the establishment of a prognosis-behaviour model will have to be considered in future.

With the implementation of the above mentioned first step of a "Pavement Management System" for the Namibian Department of Transport, data about roads in the Namibian paved road network have easily become available. This allows analysis to be done on the data quicker and easier than in the past. One of the objectives of such analysis is the development of a predicting behaviour model for road pavements. Such a predicting model has been developed [51] using data from a Brazil pavement performance study.

The predicted change in roughness is described as follows:

dR=0,563*(e3,0*DEF*MR1,23/SNC3,46)*dNe + 0,90dSPA+0,45dDPA+2,94dRDS

dR = Change in roughness (QI) (counts/km) (IRI=(QI+10)/14) [37]
DEF = Benkelman Beam Deflection (mm)
MR = Mean roughness (QI)
dNe = Change in cumulative number of equivalent 80 kN axle loads per lane in millions
SNC = Corrected pavement structural number
dSPA = Change in percentage area of surface patching
dDPA = Change in percentage area of deep patching
dRDS = Change in standard deviation of lane rut depth (mm)

This relationship has the form:

Increase in roughness = Roughening due to traffic + roughening due to other causes

It is proposed for simplifying reasons under Namibian conditions to use only the roughening due to traffic. The equation can thus be integrated to calculate the predicted remaining life in terms of E80 axles:

Ne = (SNC3,46/0,563*e(3,0*DEF))*(Rp(-0,23)-Rt(-0,23)/0,23) (E80*106)

where Rp is the present roughness and Rt is the terminal roughness in QI (counts/km) [52].

This equation is used to predict the remaining life of a pavement and can be regarded as "Performance Ratio", but cannot be related directly to the performance of a specific pavement structure due to lack of data. In Namibia, for instance, no deflection measurements are available yet. To bridge this problem the deflection measurements are estimated [53]. This formula has to be used with reservation, but due to lack of better relationships it has to be used for the time being:

DEF = 2,427*SNC(-1,072)

Above performance prediction model establishes the incremental change in riding quality from the stage of the present riding quality and takes the present riding quality into account when predicting remaining life.

To establish the Corrected Pavement Structural Number 'SNC' which takes the subgrade strength into account the following equation is used [54]:

SNC = SNU+3,51 log10CBR-0,85(log10CBR)2-1,43

where:

SNU = Eall layersS.C.i*ti/25 Structural Number for layers
S.C.i = layer strength coefficient according to NOTA to table 7 [91]
ti = layer thickness (mm)
CBR = California Bearing Ratio of the subgrade (table 7)
If X = 3,51 log10CBR-0,85*(log10CBR)2-1,43 then:
X = 1,23 for CBR = 10; X = 1,52 for CBR = 15; X = 1,70 for CBR = 20;
X = 1,82 for CBR = 25 and X = 1,90 for CBR = 30

Thus: SNC varies between 3,0 and 4,5 for average conditions

With DEF = 2,427*SNC(-1,072):
DEF = 0,63 mm for SNC = 3,5 and DEF = 0,55 mm for SNC = 4,0

The validity of the prediction model was checked [55] by assessing the accuracy of the past performance of some Namibian pavements. These were compared against the model. An average initial riding quality of IRI=2,5 (m/km) on a new road was assumed and an average terminal riding quality of IRI=3,7 (m/km). These theoretical established riding quality predictions were compared with actual riding quality measurements. The scatter around the theoretical prediction line for different ratios of different past traffics to life traffics gave an indication of the accuracy of the model. The scatter can be attributed to the variation of the initial riding qualities. The riding quality predicted by the model has been proved to rarely deviate more than IRI=0,5 (m/km) on pavements varying in ages from 3 to 20 years. However, a IRI=0,5 (m/km) is equivalent to about 40% of the life of a pavement. It can thus be assumed that the estimates using this model can be expected to have an error of about 40% of the predicted life.

Due to the relative inaccuracy of the prediction model with an assumed error of 40% it is unrealistic to compute the exact number of remaining years of life of a pavement but rather to establish the approximate number of E80 axles a pavement can be expected to carry until it reaches its terminal value of a IRI=3,7 (m/km).

Analysis of the investigations to establish predicted remaining life of pavements (see also next section and table 32) proved that the pavement structures of Namibia's paved roads are sound and in many cases economically not appropriate if the past cumulative traffic in terms of E80 axles is compared with the E80 axle traffic which the roads can carry until a terminal riding quality of IRI=3,7 is reached, as long as the surfacing is kept in good order to keep traffic loads and water from the basecourse and other pavement layers. Most roads can carry considerable more E80 axles before the terminal IRI value will be reached. In many cases it is much more probable that capacity problems will arise before a terminal IRI value of 3,7 will be reached (example: T.R.6/1: Windhoek-Airport). If higher IRI values, as for instance a IRI value of 4,5 or even 5,5 will be accepted, the life time of most Namibian paved roads will increase even further. It must, however, be kept in mind that the present IRI values are only average mean values over very long distances of road (see table 32). In many cases higher IRI values will occur on short sections (some 100 metres to some kilometres), with the subsequent smaller amounts of E80 axles which the pavement can carry before the end of the serviceability of that section of the pavement will be reached. In these isolated cases earlier rehabilitation due to serviceability failures of the pavement structure can be expected. The lesson learnt, are that the material specification for the layerwork structure of the pavement are currently too high and can be relaxed. For instance, optimum-moisture- CBRs can be used in place of soaked CBRs as mentioned earlier, and the results of test sections where relaxed materials specifications have already been successfully applied, can be incorporated into a future relaxed standard specification for Namibian paved roads. Much room for the relaxation of standards is evident especially in the case of the currently too stringent specifications for gradings and plasticity indexes ' PI'. The proviso must be that routine maintenance and early reseals, as outlined above, must timeously applied to the surfacing in order to protect the pavement structure from penetrating water. For some sections of paved roads which needed complete rehabilitation after an approximate lifetime of 20 years, like for instance some sections of trunk road 1/7 between Okahandja and Sukses or some sections of trunk road 1/3 between Keetmanshoop and Mariental, it was not the end of duration of life of the pavement, the reaching of the terminal value of E80 axles or materials fatigue, but rather construction errors with specification violations attended with poor roadbeds and road drainage as well as penetrating water which resulted in pavement serviceability failures.

 

6.7  THE RELATIONSHIP BETWEEN 'VOC' AND 'PMS'

 

As the Namibian "Pavement Management System" for the determination of rehabilitation projects is based on the ' PMS' system of the Provincial Road Authority of the South African Cape of the Good Hope, the " Net Present Worth" of rehabilitation 'W' can be calculated as follows [56]:

W = VOCs + MCs - RCr (US $/km/year)

where: VOCs = savings in road user costs (based on riding quality and volume of traffic
MCs = savings in routine maintenance costs
RCr = annual capital recovery costs of rehabilitation

The savings in routine maintenance costs and the annual capital recovery costs of rehabilitation of a road project can be calculated with above formula. The user savings are calculated on the basis of a relationship developed by the RSA/Namibian 'VOC' study based on the figures computed in tables 25, 26 and 27 for the year 1989. All vehicle operating costs are unfactorised net costs without any government taxes and subsidies. During the first South African 'VOC' investigations during 1982 it was empirically assumed that for a riding quality better than IRI=3,3 there are no vehicle operating costs benefits. This value of IRI=3,3 was therefore accepted as cut-off point for the calculation of 'VOC' savings and all savings for a riding quality of less than IRI=3,3 were calculated for the three vehicle classes dealt with in tables 25, 26 and 27, namely passenger busses, passenger cars and medium trucks.

The 1985 South African 'VOC' study used a standardised traffic volume expressed in equivalent vehicle units (evu), where one truck is equivalent to three passenger cars and one passenger car is equivalent to one 'evu'. The 'evu' is a measure for the capacity and the cost implications of the performance of a specific road and not one for the traffic load which is measured in terms of E80 unit axles. A traffic stream consisting of 15% heavy vehicles was further assumed in above 'VOC' study. Based on this assumption which is realistic for Namibian conditions as an average, the weighted savings per 'evu' were then established. A second order equation was fitted to the annual savings per 'evu' and kilometre, and the following model resulted from the RSA/Namibian ' VOC' study:

VOCs = 0,5209*(55,3 - 34,6*e(1,63-0,0173*(14*IRI-10)) + 5,26*(e(1,63-0,0173*(14*IRI-10)))2)
(US $/evu*km*year)

This model had a standard error of estimate of 0,6 and its validity was established for the range 3,3<IRI<7,5 due to the fact that to date no paved road had been encountered in Namibia with a riding quality worse than IRI=7,5. In actual fact the worst average riding quality so far established on the total Namibian paved roads system was only IRI=4,5 which of course doesn't mean that there couldn't be isolated short sections with IRI values of more than this figure. The 'VOC' model for Namibian roads and vehicle operating cost savings is reflected in the HP 1000 computer program '&RR705' of the Namibian Department of Transport which evaluates rehabilitation projects of paved road sections. It must, however, be observed that the total 'VOC' savings are not very significant in terms of money, due to the very low traffic numbers in terms of 'evu' on Namibian roads. Table 32 is based on the calculations of remaining life of pavements and the savings in US $ per 'evu', kilometre and year will be calculated for the different average IRI values (for IRI values > 3,3 only) for all Namibian paved roads with an average inflation rate of 75% for vehicle capital and running costs for the period between 1985 and 1989:

 

TABLE 32  PREDICTED SAVINGS IN 'VOC' FOR PAVED ROADS

 

|==================================================================|
|
ROAD    | START  | END  | PRESENT | GR. | SNC | IRI | Y.| PAST | PREDICTED       |
|  NO     | (km)   | (km) | E80's   | RAT |     |  *  | OF| CUMUL| SAVINGS IN VOC   |
|         |        |      | P.LANE  |  %  |     |     | OP| E80s | BASED ON IRI     |
|         |        |      | P.DAY   |  ** |     |     | EN| x106 | $/evu*km*year    |
|---------|--------|------|---------|-----|-----|-----|---|------|------------------|
|TR0101+  |   0,00 |142,50|    45   |  2  | 3,88|  2,2| 68| 0,251|        0         |
|TR0102+  |   0,00 |152,62|   155   |  4  | 3,63|  2,4| 67| 0,773|        0         |
|TR0103+  |   0,00 | 53,42|   190   |  4  |  ND |  2,8| 61| 1,127|        0         |
|         |  53,42 |129,75|   190   |  4  | 3,08|  2,8| 66| 0,980|        0         |
|         | 129,75 |228,71|   200   |  4  |  ND |  2,9| 63| 1,128|        0         |
|TR0104+  |   0,00 | 15,35|   198   |  4  |  ND |  3,2| 62| 1,146|        0         |
|         |  15,35 |174,86|   190   |  4  |  ND |  2,9| 63| 1,071|        0         |
|TR0105+  |   0,00 | 86,49|   250   |  4  |  ND |  3,0| 61| 1,482|        0         |
|TR0106+  |   0,00 | 21,88|   250   |  6  |  ND |  2,2| 78| 0,601|        0         |
|         |  21,88 | 68,21|   300   |  4  |  ND |  2,1| 58| 1,900|        0         |
|TR0107+  |   2,04 |120,00|   200   |  5  | 3,22|  3,2| 64| 1,010|        0         |
|         | 120,00 |173,66|   200   |  5  | 3,59|  2,7| 65| 0,983|        0         |
|TR0108+  |   0,00 | 39,40|   187   |  5  | 3,17|  3,0| 66| 0,893|        0         |
|         |  39,40 |118,76|   150   |  5  | 3,17|  3,0| 66| 0,716|        0         |
|TR0109+  |   0,00 | 61,80|    50   |  4  |  ND |  2,9| 60| 0,303|        0         |
|TR0110+  |   0,00 | 94,75|   225   |  6  | 5,10|  1,9| 67| 0,971|        0         |
|TR0111+  |   0,00 |155,22|   225   |  6  | 5,10|  2,2| 68| 0,943|        0         |
|         | 155,22 |194,94|    30   |  3  |  ND |  2,3| 85| 0,022|        0         |
|TR0201+  |   0,00 | 30,40|   105   |  4  |  ND |  2,4| 68| 0,505|        0         |
|TR0202+  |   0,00 | 54,79|   125   |  6  | 3,69|  2,4| 66| 0,555|        0         |
|         |  54,79 | 67,82|    35   |  4  | 3,48|  2,5| 67| 0,174|        0         |
|         |  67,82 |144,00|    35   |  4  | 3,11|  2,2| 65| 0,186|        0         |
|TR0203+  |   0,00 | 33,40|    35   |  4  | 3,32|  2,4| 66| 0,181|        0         |
|         |   33,40| 95,43|    12   |  4  | 3,56|  2,3| 77| 0,034|        0         |
|TR0204+  |    0,00| 44,15|     7   |  4  | 3,24|  2,3| 74| 0,025|        0         |
|         |   44,15| 99,86|     7   |  4  | 4,38|  2,2| 73| 0,026|        0         |
|         |   99,86|132,68|    10   |  4  | 4,38|  2,2| 72| 0,040|        0         |
|TR0205+  |    0,00| 10,52|    25   |  4  | 2,84|  2,9| 69| 0,116|        0         |
|         |   10,52| 54,81|    25   |  4  | 2,45|  2,5| 70| 0,111|        0         |
|         |   54,81| 68,90|    25   |  4  | 2,83|  3,3| 70| 0,111|        0,28      |
|TR0301+  |    0,00| 48,12|    75   |  4  | 4,14|  2,7| 68| 0,360|        0         |
|         |   48,12|100,11|    75   |  4  | 3,90|  2,7| 68| 0,360|        0         |
|         |  100,11|114,21|    75   |  4  | 3,05|  2,8| 67| 0,374|        0         |
|         |  114,21|178,08|    75   |  4  | 3,54|  2,7| 67| 0,374|        0         |
|TR0401+  |    0,00| 76,26|     9   |  4  | 3,85|  2,5| 73| 0,034|        0         |
|         |   76,26|107,17|    11   |  4  | 3,90|  2,7| 73| 0,042|        0         |
|TR0402+  |    0,00|121,76|    11   |  4  | 3,96|  2,5| 75| 0,037|        0         |
|TR0501+  |    0,00| 53,63|    22   |  4  | 5,07|  2,5| 70| 0,097|        0         |
|TR0601+  |    0,00| 15,80|    85   |  4  | 4,30|  2,2| 61| 0,504|        0         |
|         |   15,00| 37,99|    85   |  4  | 3,30|  2,2| 65| 0,453|        0         |
|         |   37,99| 90,76|    75   |  4  | 3,00|  2,8| 71| 0,316|        0         |
|         |   90,76|198,28|    52   |  4  | 3,00|  2,9| 71| 0,220|        0         |
|TR0701+  |    0,00|112,57|    50   |  4  | 3,29|  2,7| 64| 0,274|        0         |
|TR0801+  |    0,00| 90,36|   100   |  5  | 3,22|  2,4| 71| 0,398|        0         |
|TR0802+  |    0,00| 17,28|   145   |  6  | 6,12|  2,8| 70| 0,567|        0         |
|         |   17,28| 58,72|   140   |  8  | 4,53|  2,5| 83| 0,142|        0         |
|         |   58,72|127,41|   137   |  8  | 4,09|  2,7| 83| 0,139|        0         |
|TR0803+  |    0,00|131,03|   137   |  8  | 5,47|  2,5| 80| 0,250|        0         |
|TR0806+  |    0,00| 98,76|    20   | 10  |  ND |  1,6| 76| 0,049|        0         |
|         |   98,76|117,98|    13   | 10  |  ND |  1,9| 77| 0,030|        0         |
|TR1501+  |    0,00| 61,81|    15   |  4  | 3,73|  2,8| 77| 0,042|        0         |
|MR0031+  |    0,00| 31,88|     5   |  3  | 3,65|  2,8| 73| 0,020|        0         |
|MR0034+  |    0,00|109,51|     4   |  3  |  ND |  2,8| 80| 0,008|        0         |
|MR0047+  |    0,00| 11,82|    10   |  4  |  ND |  2,9| 80| 0,020|        0         |
|MR0052+  |    0,00| 16,00|    27   |  4  | 4,54|  3,3| 71| 0,114|        0,28      |
|         |   16,00| 22,00|    12   |  4  | 4,54|  3,2| 72| 0,048|        0         |
|MR0055+  |    0,00| 14,75|   105   |  4  | 3,63|  2,7| 71| 0,443|        0         |
|         |   14,75| 14,87|    40   |  4  | 3,63|  2,7| 71| 0,169|        0         |
|MR0061+  |    0,00| 77,45|    15   |  4  | 4,22|  2,3| 73| 0,057|        0         |
|MR0067+  |    0,00| 10,34|    19   |  4  |  ND |  2,8| 70| 0,084|        0         |
|         |   10,34| 50,69|    19   |  4  |  ND |  2,4| 79| 0,043|        0         |
|         |   50,69| 82,86|    15   |  4  |  ND |  2,5| 80| 0,030|        0         |
|         |   82,86|155,60|    11   |  4  |  ND |  2,4| 81| 0,019|        0         |
|MR0068+  |    0,00| 50,00|     7   |  3  |  ND |  2,2| 85| 0,005|        0         |
|MR0070+  |    0,00| 14,82|    20   |  4  | 3,14|  3,0| 75| 0,067|        0         |
|MR0072+  |    0,00| 55,80|   150   |  6  | 3,90|  2,8| 71| 0,564|        0         |
|MR0084+  |    0,00| 35,55|     3   |  3  | 4,08|  2,4| 69| 0,015|        0         |
|MR0087+  |    0,00| 26,24|     5   |  3  | 2,71|  1,6| 81| 0,009|        0         |
|MR0092+  |    0,00| 34,46|   225   |  6  |  ND |  2,5| 69| 0,912|        0         |
|         |   34,46| 34,68|   150   |  6  |  ND |  3,2| 85| 0,106|        0         |
|         |   34,68| 36,42|   120   |  6  |  ND |  3,5| 85| 0,085|        1,24      |
|         |   36,42| 60,75|    75   |  6  |  ND |  3,0| 84| 0,053|        0         |
|         |   60,75| 85,16|    45   |  6  |  ND |  3,2| 83| 0,046|        0         |
|         |   85,16|122,73|    30   |  6  |  ND |  2,7| 83| 0,031|        0         |
|         |  122,73|161,93|    23   |  6  |  ND |  2,7| 82| 0,031|        0         |
|         |  161,93|186,75|    15   |  6  |  ND |  2,8| 82| 0,020|        0         |
|         |  186,75|208,00|    11   |  6  |  ND |  2,7| 68| 0,046|        0         |
|MR0101+  |    0,00| 39,60|    16   |  4  | 3,87|  2,8| 81| 0,027|        0         |
|         |   39,60| 72,13|    15   |  4  | 4,03|  2,7| 81| 0,025|        0         |
|DR1491+  |    0,00| 11,72|     8   |  3  | 3,76|  2,3| 78| 0,021|        0         |
|===================================================================================|
NOTA: Table 32 has been computed using data from table 7 as well as above mentioned equation. Prices for December 1989.
* IRI measured 1984/86; ** Growth rates: E80 axles: estimates

Table 32 shows clearly that, as assumed above, the Namibian paved roads system possesses very high average riding qualities in terms of IRI with the result that the savings in road user costs to achieve a maximum of IRI=3,3 are minimal. It can also be derived that vehicle operating costs have not any meaningful influence on the establishment of a rehabilitation algorithm as represented in the pavement management computer program of the Namibian Department of Transport '&RR705' due to the low traffic numbers. It can be further concluded that any rehabilitation of surfaced pavements can only be justified by direct routine maintenance costs as, for instance, the costs of " Bitumen Maintenance Units" but not by vehicle operating costs alone. A high IRI value of, say, between 7,00 and 8,00 for a paved road is surely dangerous and will have to be adjusted by rehabilitation, even if it cannot be justified by the economics of ' VOC'. Thus it can be established that under Namibian conditions rehabilitation of surfaced road pavements will have to be a function of poor riding quality, bad rutting, skid resistance and high routine maintenance expenditure but not one of vehicle operating costs. It could not be proved that where rehabilitation of surfaced road pavements could not be justified by 'VOC', timing and accident costs would be decisive factors. Empirically it was found that a riding quality of IRI=4,5 can be regarded as cut-off point for economic rehabilitation of a paved road. This is the reason that 'VOC' were not included into the Namibian ' PMS' maintenance system for paved roads. The picture will be different when the 'VOC' principle will be applied to unpaved roads. The " Maintenance Design System" for unpaved roads is based on 'VOC' on the basis of the for southern Africa adapted HDM3 maintenance system.

 

6.8  PROPOSAL FOR A RELAXED MATERIALS SPECIFICATION

 

The evaluation of the data of all pavement structures of Namibian paved roads and the application of a performance model based on these data brought to light that most of the pavements have been not appropriate in an economic sense. Using the predictions resulting from this performance model, it can be concluded that many so far designed and constructed pavement structures have carried to date only a small fraction of E80 axles which such pavements could carry in terms of the total number of E80 axles before reaching a serviceability failure in form of the terminal value of the IRI.

It is proposed for the future road network in Namibia to apply new ideas to create revised specifications and design standards for rural roads. Considerable savings can be ascertained by the establishment of a revised appropriate composition of pavement layers.

Such appropriate specifications will not so much concentrate on the relaxation of CBR values or densities but rather on the reduction of the number of selected pavement layers and their thicknesses. The CBR value is a yardstick for the strength of a pavement layer and any volumetric change of a pavement layer under traffic, caused by sub-standard CBRs, will have to be avoided. But the practice followed currently by the Namibian Department of Transport to determine CBR values of the subgrade soils carried out on specimens soaked for several days before test, has to be relaxed. Above practice leads invariably to an under-estimation of the field strength of some soils and hence can lead to a more expensive design than is necessary. Especially under the prevailing climatic conditions in Namibia the subgrade rarely ever approach the soaked conditions. Tests under soaked condition are only applicable to subbase or subgrade materials beneath the basecourse which are highly permeable and have high void contents like for instance stone pitch fills which are seldom used in Namibia, or under circumstances, where it is apparent that poor quality work is all that could be expected. But, this has not yet been proved under Namibian conditions.

Further relaxations can be applied in the acceptance of higher PI values and less stringent specifications for the grading requirements of specific selected pavement layers, especially for pedocrete layers [57]. Experience showed that grading is not one of the most important quality assuring properties of a CBR value under Namibian conditions. In the past decades a lot of research has been done in this regard in the Central Laboratory of the Department of Transport and it has been proved that in the judgement of CBR values grading is not so important as was assumed earlier. If the plasticity and grading requirements can be relaxed, expensive equipment in the physical modification of basecourse or subbase pavement layers, like crushing equipment and rock busters, can be eliminated and haul distances reduced. Expensive processing of pavement layers could have been avoided in the past if the specifications regarding plasticity, gradings and soaked CBR values would have been less stringent.

The proposed materials specification for paved roads will incorporate the following relaxations for normal Namibian materials and traffic conditions:

- Reduce number of selected layers from 4 (even 5) to 3 (1 selected subgrade layer is regarded sufficient)
- Reduce thickness of selected layers from 150 (respective 125 mm) to 100 mm
- Reduce the current minimum Grading Modulus ' GM' from 1,7 (respective 2,0-2,1 average) to 1,5
        GM =[(300 - E of % passing 2,0; 0,425 and 0,075 mm sieves)/100]
- Increase the permissible Plasticity Index ' PI' for: basecourse: from 6 (respective 8 or 10) to 15
        subbase: from 10 (respective 12 or 15) to 20
        selected layer: from 15 (respective 20) to 25
        subgrade: from 20 (respective 25) to 30
- Minimum CBR will be retained for:
        basecourse: 80 (120 stabilised)(soaked)(98% Mod AASHO)
        subbase: 45 ( 70 stabilised)(optimum moisture)(95% Comp)
        selected layer: 10 (optimum moisture)(93 Mod AASHO)
        subgrade: 3 (optimum moisture)(90 Mod AASHO)

If with above relaxed materials specifications for pavement layers a minimum CBR of 80 (for basecourse) can be achieved a sound pavement structure can still be maintained, with a considerable reduction of costs against the current materials specification. With the new proposed specification vast savings can be achieved. Thus, the prospects for significant advances in the direction of optimised construction and maintenance models for paved roads are favourable in Namibia.

WB00823_.GIF (134 bytes)

[Return to Contents]

forward.GIF (132 bytes)