Highway Research Bulletin

YEAR 2004-2005
Bulletin No. 73

A rational approach for design of flexible overlays for Indian highways
By M. Amaranatha Reddy*, K. Sudhakar Reddy** & B.B. Pandey***


ABSTRACT

The current Indian Roads Congress Specifications1 for design of overlays for flexible pavements recommend the use of Benkelman beam rebound deflection measurement technique for evaluation of the structural condition of in-service flexible pavements. Though Benkelman beam technique is simple and has been adopted by various agencies throughout the world, it is being gradually replaced by more advanced approaches. Advancements made in the area of instrumentation made it possible for more realistic and rational evaluation of pavements. Falling Weight Deflectometer (FWD) is one such equipment that is being used routinely by many agencies for the structural evaluation of pavements. A number of agencies have developed procedures for design of overlays that require evaluation of pavements using FWD. This paper is an outcome of the research work carried out by the transportation engineering section of the Indian Institute of Technology, Kharagpur with funding from the Ministry of Road Transport and Highways (MORTH), Government of India. The paper presents an FWD based procedure for design of overlays for flexible pavements in India. Salient features of an In-vehicle FWD and a genetic algorithm based backcalculation procedure are discussed. Details of investigations carried out on different highways in India are given. An FWD based overlay design procedure developed on the basis of these investigations is presented in the Paper.
--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
* Senior Lecturer Department of Civil Engineering, National Institute of Technology, Hamirpur - 177005, Himachal Pradesh, India.
** Professor Department of Civil Engineering, Indian Institute of Technology, New Delhi - 110016, India.
--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
1. INTRODUCTION
Highway sector in India has started getting major attention in the recent past with the initiation of several major highway projects by the central and state governments. Many of the existing national and state highways are being widened and strengthened. Several new highways and expressways are being constructed to handle the demand placed on the system by the fast growing commercial traffic volumes and the ever-increasing axle loads of commercial vehicles. The National Highway Development Project (NHDP), a major highway development project, currently under implementation in India, includes Golden-quadrilateral scheme, and East-West and North-South corridor projects. This huge network of road system has to be evaluated regularly to assess its structural and functional condition using suitable evaluation techniques.

With increase in the demand for efficient road network, awareness to use rational or mechanistic pavement and overlay design procedures is also increasing among highway professionals in India. Over the last one decade, a gradual shift, from empirical to mechanistic approach, has been seen in India in the methodology of pavement design. Research efforts made during the last few decades resulted in the development of rational design methods for pavement design as well as overlay design. The recent revision of Indian Roads Congress Guidelines2 for design of flexible pavements adopt mechanistic performance criteria. However, Benkelman beam rebound deflection technique1 is still used in India for the design of overlays for flexible pavements. In this technique, a single rebound surface deflection is measured under a static load, which is not representative of the short duration load pulses produced by vehicles on highways. Assessment of structural capacity of the pavement on the basis of a single deflection does not yield adequate information about the condition of various layers in the pavement. The fact that evaluation of pavements using Benkelman beam is a very slow process also makes it an unattractive option.

This paper is the outcome of the research work carried out by the Transportation Engineering section of the Indian Institute of Technology (IIT), Kharagpur for the development of an FWD based pavement evaluation and overlay design procedure for Indian highways. The work was carried out with financial assistance from the Ministry of Road Transport and Highways, Government of India through a Research Project R-81 Structural Evaluation of Pavements using Falling Weight Deflectometer. The work included (i) fabrication of an in-vehicle falling weight Deflectometer (ii) development of a genetic algorithm based software for backcalculation of pavement layer moduli (iii) structural evaluation of in-service and new technology pavements using the FWD and (iv) recommendation of a rational approach for design of overlays for flexible pavements. Salient features of these developments/investigations are presented in this paper.

2. STRUCTURAL EVALUATION OF PAVEMENTS USING FWD
Different types of equipment are used for evaluating the structural condition of pavements. These equipment differ from one another in terms of the applied peak load and loading pattern. The structural responses commonly measured are deflections at a number of locations on the pavement surface. The measured deflections are used to evaluate the structural condition of the in-service pavement. Among the various equipment available for structural evaluation of pavements, Falling Weight Deflectometer (FWD) is extensively used worldwide because it simulates, to a large extent, the actual traffic loading conditions. Deflections measured with FWD are used for backcalculating the moduli of the pavement layers. The computed moduli are, in turn, used for (i) the evaluation of different layers of in-service pavements (ii) the estimation of their remaining lives and (iii) determination of strengthening requirement, if any.

FWD is an impulse-loading device in which a transient load is applied to the pavement and the deflected shape of the pavement surface is measured. The duration of impulse is approximately equal to the time needed to traverse the length of a tyre imprint at a speed of about 60kmph3. The working principle of a typical FWD is illustrated in Fig. 1. Impulse load is applied by means of a falling mass, which is allowed to drop on a spring placed over a loading plate. The deflected shape of the pavement surface is measured using velocity transducers (geophones) placed at different radial distances from the center of the load. Characteristics of the deflection bowl along with other pavement parameters are used to backcalculate the effective layer moduli using an appropriate backcalculation technique.



3. FWD EVALUATION BASED OVERLAY DESIGN PROCEDURES
The purpose of overlay design is to determine the requirement of overlay, which, when placed over the existing pavement, will restore its structural condition so that the pavement performs satisfactorily during its design period. The key element in any overlaydesign procedure is the system adopted for the evaluation of the structural condition of the in-service pavement. Many of the recently developed overlay design procedures recommend pavement evaluation using FWD.

Brown et al,4 proposed an approach for overlay design based on FWD deflection measurements. In this method, FWD deflection studies are conducted at two places, i.e, at the center of the lane and along the wheel track. These deflections are used to backcalculate the layer stiffnesses, which are then adjusted for temperature and speed. The performance of the pavement in fatigue cracking and permanent deformation modes of failure is estimated using Nottingham performance criteria5. The back calculated layer moduli obtained using center lane deflections are used to determine the original fatigue, life of the pavement. Fatigue damage is computed as :

Fatigue damage = Np/ Nt ... Equn. 1

where, Np = traffic carried by the existing pavement in msa; Nt = estimated fatigue life of the existing pavement, msa.
Overlay thickness requirement for extending the fatigue life is obtained using Miner’s principle in which the total damage must be less than or equal to 1. If the design traffic is ‘N’ then

Np/ Nt + N / Nn = 1 ... Equn. 2

where, Nn = new total fatigue life of the pavement at the reduced level (due to overlay) of tensile strain.

Thickness requirement from permanent deformation consideration is also determined in the same manner. Using the backcalculated moduli adjusted to the design conditions, maximum vertical subgrade strain at the top of the subgrade is calculated. The overlay thickness requirement is obtained from a plot of traffic-induced versus thickness. The larger of the two thicknesses obtained from fatigue and rutting criteria is chosen as the overlay thickness.

PAVMAN computer program was developed by Richer and Irwin6 to determine overlay thickness requirement and remaining life of the existing pavement. MODCOMP 2 was used to determine the pavement layer moduli from FWD data.

Mamlouk7 proposed a rational overlay design method for flexible pavements in Arizona State based on roughness, fatigue and plastic deformation models. FWD tests were conducted and the backcalculated moduli values were used to develop fatigue and plastic deformation models. These models were incorporated in the microcomputer program CODA, which calculates overlay design thickness and the remaining life of the pavement.





4. CURRENT INDIAN PRACTICE FOR DESIGN OF FLEXIBLE OVERLAYS
In the Indian Roads Congress method1, overlay design thickness is estimated using Benkelman Beam deflection data. The measured deflections are corrected for a standard temperature of (35oC) and for seasonal effect. Design charts are available in the Indian Roads Congress guidelines for estimating the overlay thickness in terms of bituminous macadam layer based on design traffic and characteristic deflection. While the structural evaluation of pavements using Benkelman beam is simple, the fact that it does not simulate the dynamic loading conditions that normally prevail on highways makes this approach unsuitable for rational evaluation of flexible pavements. The single rebound deflection measured in this method is not likely to reveal a great deal of information about the condition of the individual layers. In view of these reasons and considering that the Indian Roads Congress has recently adopted mechanistic performance criteria for design of flexible pavements2, the Ministry of Road Transport and Highways, Government of India initiated a major research project (R-81) for structural evaluation of pavements using FWD in the transportation engineering section of Indian Institute of Technology, Kharagpur. The main objectives of the project were the development of a pavement evaluation procedure using a low-cost in-vehicle falling weight deflectometer developed in India. Details of the FWD based procedure developed for design of overlays for flexible pavements in India are given in the following sections.

5. DEVELOPMENT OF AN FWD-BASED OVERLAY DESIGN
PROCEDURE FOR INDIAN HIGHWAYS
The main components of the research work conducted by IIT, Kharagpur for the Ministry of Road Transport and Highways, Government of India are:-
1. Development of an indigenous in-vehicle falling weight Deflectometer
2. Development of genetic algorithm based software for backcalculation of pavement layer moduli
3. Evaluation of in-service pavements using FWD
4. Development of fatigue and rutting performance criteria
5. Formulation of design steps for design of overlays.

5.1. IIT Kharagpur FWD (IITKGP FWD)

IITKGPFWD16 is an in-vehicle falling weight Deflectometer developed by the Transportation Engineering group of IIT, Kharagpur. The in-vehicle assembly allows easy maneuvering of the FWD unit on heavily trafficked Indian two-lane two-way highways. It has facility to vary both the falling mass (100 to 225 Kg) and the height of fall (100 to 600 mm). Impulse load in the range of 20 kN to 100 kN can be obtained by selection of suitable mass and height. A loading plate of 300 mm diameter is used. Larger loading plates also can be used. Load pulse duration of 20 to 30 ms is attained.

Velocity sensors (geophones), housed on a geophone frame, are used to measure surface deflections. The geophone frame has been made foldable so that it projects beyond the rear of the vehicle only during testing. Once the test is over, the frame is folded back. Operations such as raising and lowering of mass, movement of the geophone frame and catching and releasing the load are controlled with the help of push buttons. These operations can also be controlled in automated mode with the help of an on-board Programmable Logic Circuit (PLC). A set of batteries is used to supply power to computer, data acquisition system and PLC. Power for operating hydraulic system is drawn from the engine of the vehicle. Load and deflection data are acquired with the help of a high-speed data acquisition system. Photo 1 shows a view of the FWD.




Photo 1. Evaluation of National Highway-6 using In-vehicle FWD



5.2. GA based Backcalculation Program
A Genetic Algorithm (GA) based program, BACKGAIIT, was developed for backcalculation of effective moduli of pavement layers17.18. When used as inputs to elastic layered theory, effective layer moduli produce responses (deflections) similar to those observed in field. ELAYER program19 developed for analysis of layered pavement systems was used for forward calculation of surface deflections in BACKGAIIT.

Inputs to BACKGAIIT program include:- measured deflections, radial distances at which deflections are measured, layer thicknesses, Poisson ratio values of different layers, applied load and loading plate radius. It is also necessary to give practical ranges for different layer moduli to be estimated.

5.3. Evaluation of Pavements Using IITKGP FWD
Various types of in-service and some new technology pavements with different thicknesses of bituminous surfacing and granular base were selected for detailed investigation. These sections are located in different parts of the states of West Bengal (WB), Orissa (OR) and Jharkhand (JH) in eastern India. Most of the in-service pavements have two-lane undivided carriageway and new pavements have four-lane divided carriageway. Average daily two-way traffic on these roads ranged from 300 to 7000 commercial vehicles per day (cvpd).

5.3.1. In-service Pavement Sections : In-service pavements consisted of thin layer bituminous surfacing applied from time to time over granular subbase and base. The base and subbase consisted of layers of sand, brickbat and crushed stone aggregates in varying thicknesses. All the granular layers in a pavement section were treated as a single layer (granular base) for analysis. Similarly, the bituminous surfacing layer consisted mostly of bituminous macadam covered with premix carpet and seal coat. One or more layers of bituminous material placed over the granular layer at different times with varied thicknesses were also considered as one layer. It was observed at the time of investigation that some of the pavement sections were badly cracked and some were showing cracks covering nearly 20 per cent of the pavement area.

5.3.2. New Pavement Sections : Some of the newly constructed pavement sections on NH-6, which formed part of Golden quadrilateral scheme of the National Highway Development Programme (NHDP) for connecting the major metropolitan cities of India, were selected for FWD evaluation during different stages of construction. The pavement consisted of four-lane divided carriageway. The lengths of the selected test stretches varied from 300 m to 800 m. A typical layout of test section is shown in Fig. 2.

2 Lane 2 way
carriageway
4 Lane Divided
Carriageway

Details of the selected test sections, both in-service and new pavements, are given in Table 1. Photo 2 (a) and 2 (b) show different views of the FWD being used for structural evaluation of pavements.




a) General View




b) Geophone Frame Being Lowered to the Pavement Surface



Deflections were measured during three seasons of the year, i.e. monsoon (M), winter (W) and summer (M) seasons. The deflections for monsoon season were measured immediately after the monsoon receded. In addition to the deflection data, air and pavement temperatures were also recorded at each test location. Figs. 3 and 4 present typical deflection data measured during three different seasons on two different pavement sections.


A limited study was conducted to evaluate the effect of pavement temperature on measured deflections and layer moduli. This study was conducted at km 112, 126 and km 131 (stretch Nos. 19, 21 and 22 of Table 1) of National Highway No.6. Deflections were measured at each test location at different time intervals as the pavement temperature was varying. Deflection measurements were taken at 10 locations in Km 130 .00 to 131.10. Fig. 5 shows the variation of deflection with temperature for Km 130.100, which is average value of three trials.


5.4. Backcalculation of Layer Moduli Using BACKGAIIT Program
The structural behaviour of a pavement in mechanistic analysis is indicated by the elastic modulus and Poisson’s ratio of different layers. Effective modulus values are backcalculated in the present study from measured deflections. In the context of backcalculation, effective modulus, is the representative modulus of a layer which produces the same response as obtained by FWD. Deflection data obtained using the FWD was used to backcalculate effective moduli of pavement layers using BACKGAIIT program. Salient features of BACKGAIIT have been discussed in the previous section. The pavement sections were considered as three-layer elastic systems consisting of bituminous surfacing, granular base and subgrade. The inputs required for backcalculation analysis are the thicknesses of the first two layers and Poisson ratio values of all the three layers. Thicknesses measured by excavating test pits were used in the analysis. Since the moduli of granular bases and subbases are not much different, the two layers were considered as a single granular layer termed as Granular Base (GB). Similarly, the thicknesses of different layers of bituminous materials were added for getting the surface thickness. Poisson ratio values of bituminous layer, granular layer and subgrade were taken as 0.5, 0.4 and 0.4 respectively. The moduli ranges considered in the backcalculation process for different situations are given in Table 2.



The surface loading considered for analysis is 40kN acting over a circular contract area with a radius of 150mm. Surface deflections measured at radial distances of 0, 300, 600, 900, 1200, 1500 and 1800mm are the main inputs to BACKGAIIT. These deflections were normalized to correspond to a load of 40 kN. The following GA parameters were used for the analysis20. Population Size = 60; Maximum number of Generations = 60; Probability of Crossover = 0.74; Probability of Mutation = 0.1. The flow chart showing various steps for the computation of layer moduli using backcalculation technique has been presented else where21.

5.4.1. Layer Moduli of Thin in-service bituminous Pavements : Pavement sections with thickness of bituminous surfacing less than 75 mm were considered as thin pavements (Thin PC) in this study. Table 3 show the details of backcalculated pavement layer moduli for pavements with thin bituminous surfacing at Km 1.825 of SH. Root Mean Square Error (RMSE) values computed using Equation 12 are given in Table.
... Equn. 12

where, DI and dI = measured and computed deflections at ith sensor, respectively;
n= number of sensors.



Similarly, for other thin bituminous pavements, the moduli values were backcalculated using the deflection data collected and other input parameters. It can be seen that the modulus value of the bituminous layer is close to that of the granular base as bituminous layer is built up of several thin layers of premix carpets laid periodically. These thin bituminous layers were mostly in a cracked condition and hence their moduli were not much different from that of the granular base layer.

5.4.2. Layer Moduli of Thick In-Service Bituminous Pavements : The pavement sections considered in the present study included a number of thick pavements with bituminous layer thickness more than 75 mm. The bituminous surface consisted of repeated application of bituminous macadam. Table 4 gives the backcalculated layer moduli for pavement section from Km 123.795 to 124.000 of NH-6. Similarly, moduli values are backcalculated for all other pavement sections considered. Thickness of bituminous layer ranged from 90 to 200 mm. A few sections had new bituminous concrete overlays of 50 mm thickness. The thicknesses granular base varied from 300 to 690 mm.



5.4.3. Layer Moduli of New Technology Pavements : The equipment used and the specifications and quality control measures being adopted currently for the construction of highway pavements in India are considerably superior to those used in past. To get an idea of moduli of pavement layers constructed as per latest Indian practices22 new pavements on some stretches of National Highways were evaluated using the FWD.
The backcalculated layer moduli for granular layer consisting of 250 mm Wet Mix Macadam (WMM), granular subbase and drainage layer of 250 mm each and two layers of Dense Bituminous Macadam (DBM), each of 85 mm thick, are given in Table 5. The data were obtained by conducting FWD tests immediately after the construction of each layer.
It can be observed from Table 5, that the subgrade modulus decreased when the FWD test was conducted after the construction of one layer of DBM (85 mm thick) and further decreased when the test was conducted on two layers of DBM (total thickness 170 mm). Similar behavior was also observed in case of base course. With high thickness of bituminous layers above the sandy subgrade and granular base course, the sum of the principal stresses due to applied load becomes smaller resulting in lower modulus values.

5.4.4. Typical Modulus Values obtained for In-service and New pavement sections: From the evaluation of several in-service pavements having thin as well as thick bituminous pavements on different highways, it is noted that the backcalculated moduli of all the layers varied from season to season and from location to location depending on the climatic, pavement and other conditions. Table 6 gives typical values of pavement layer moduli for different types of in-service and new pavements.





5.4.5. Observations on Pavement layer Moduli


5.4.5.1. Thin In-service Bituminous Pavements : The moduli values of different pavement layers of thin in-service bituminous pavements are lower during post-monsoon season and were the highest during summer. This is due to the presence of moisture in the subgrade, base and weathered bituminous surfacing of thin pavements during the monsoon (rainy) season and the resulting lower moduli values. Because of the strong dependence of the strength of the upper layers of thin pavements on that of the supporting layer (subgrade), the variations in moduli of granular and bituminous layer show trends that are similar to the seasonal variation trend of the subgrade modulus. The average bituminous layer modulus in summer season was found to be higher than that observed in winter, though bituminous material is expected to have a lower modulus value during summer. This can be explained from the fact that the bituminous layer was weathered and cracked and gave higher modulus in summer because of confinement under the loaded area coupled with absence of moisture.

5.4.5.2. Thick In-service Old Bituminous Pavements : The backcalculated subgrade modulus value was found to be higher during the summer with values ranging from 50 to 88 MPa. Similarly, granular base modulus was also found to be higher during summer (ranging from 150 MPa to 400 MPa). Lower values of moduli for subgrade and granular layers were observed during post-monsoon period. In the case of bituminous layer, there was no clear trend regarding the variation of the modulus with season. While the summer modulus is higher in some cases, in some others, the winter values were higher. This trend is due to the reason that the bituminous layers formed of repeated application of Bituminous Macadam and Pre-mix Carpet layers cracked and weathered in many locations while at some locations the bituminous surfacing did not have cracks. For uncracked bituminous layers, the winter values were higher. On the other hand, pavements with cracked surfaces behave like block pavements and there is a strong dependency of the behaviour of the surfacing on that of the underlying layers. Hence, as noted in the case of pavements with thin bituminous surfacing, summer moduli were higher for pavements with thick (cracked and weathered) bituminous surfacing.

The average values of the seasonal moduli of surface, base and subgrade of pavements with thick (old) bituminous surfacing were found to be 500, 275 and 65 MPa respectively. The average modulus value of the bituminous surface course formed due to periodical application of bituminous macadam and premix carpet is found to be 500 MPa. The subgrade is formed of laterite soil in most of the in-service pavement sections considered in the study.

5.4.5.3. New Technology Pavements : It was found that the average modulus value of Dense Bituminous Macadam (DBM) layers constructed recently on National Highways ranged from 800 MPa to 1850MPa for average pavement temperatures ranging from 25 to 400C. Average granular layer (base and subbase combined) modulus was around 280 MPa corresponding to a subgrade modulus of 60 MPa.

5.5. Effect of Temperature on Pavement Layer Moduli

Modulus of a new bituminous layer varies considerably with pavement temperature. In order to study the effect of pavement temperature on the backcalculated moduli, FWD tests were conducted on a number of sections (Sl. No.18, 20, 22 given in Table 1). The pavement temperature was recorded at a depth of 25 mm during testing at each location. The deflections were measured at temperatures of 25, 30, 35 and 40oC. Table 7 gives the layer moduli for the stretch from km 126.000 to km.126.540 on NH-6 at different pavement temperatures.



5.5.1. Development of Temperature Correction Factor : From the backcalculated moduli and the corresponding pavement temperature, an equation for estimating the factors to be used for adjusting the backcalculated moduli of bituminous layer to correspond to a temperature of 350C is obtained. The correction factor is given as :



5.6. Seasonal Variation of Pavement Layer Moduli
The backcalculated layer moduli were found to vary with season and the amount of variation is different for different types of pavements. In general, the changes in subgrade moduli are larger in different seasons compared to those observed in the case of granular and bituminous layers.

Subgrade modulus, as expected, found to be the lowest during the monsoon season. The moduli values of granular layers are found to be higher during the summer. Presence of moisture has also been found to affect the moduli of granular layers considerably. The seasonal variations in the surface moduli, though not very significant, can be attributed to the changes in pavement temperatures and to the variations in the strength of the underlying layers, especially that of subgrade. The summer moduli of bituminous layers of a number of pavement sections were found to be more than the values obtained for the other two seasons due to the effect of dry state of underlying layers.

In a given stretch of pavement, there are variations in layer thicknesses, compaction levels etc., from location to location. This is reflected in the spatial variation of the modulus values. However, among the three seasons, the scatter was found to be the maximum in the data collected in monsoon. This can be attributed to the spatial variation in the moisture content present in the pavement at different locations and the corresponding variations in the strength of the layers. Equations 13 to 21 present the relationship between the average of the moduli obtained in three different seasons and the modulus value obtained in a particular season at each location. The expressions are useful in determining the mean modulus value from the modulus backcalculated using the deflections measured during any one of the seasons.

For estimating seasonal bituminous modulus
E (bit-avg) = 15.29 (E (bit-mon) )0.5452 (R2 =0.731) ... Equn. 13

E (bit-avg) = 0.4978 (E (bit-win) ) + 207.4 (R2 =0.659) ... Equn. 14

E (bit-avg) = 0.7457 (E (bit-sum) ) + 114.69 (R2 =0.572) ... Equn. 15

where, E (bit-avg) = average seasonal bituminous modulus ; E (bit-mon), E (bit-win), E (bit-win) = moduli of bituminous layer during monsoon, winter and summer respectively (MPa).


For estimating seasonal base modulus
E (bit-avg) = 0.7224 (E (bit-mon) )+82.25 (R2 =0.5408) ... Equn. 16

E (bit-avg) = 7.623 (E (bit-win) ) 0.624 (R2 =0.685) ... Equn. 17

E (bit-avg) = 5.9244 (E (bit-sum) ) 0.654 (R2 =0.7071) ... Equn. 18

where, E (bit-avg) = average seasonal bases modulus ; E (bit-mon), E (bit-win), E (bit-win) = moduli of base layer during monsoon, winter and summer respectively (MPa).

For estimating seasonal subgrade modulus
E (bit-avg) = 0.7505 (E (bit-mon) )+ 21.69(R2 =0.438) ... Equn. 19

E (bit-avg) = 2.515 (E (bit-win) ) 0.7688 (R2 =0.808) ... Equn. 20

E (bit-avg) = 0.6420 (E (bit-sum) ) + 15.34 (R2 =0.819) ... Equn. 21

where, E (bit-avg) = average seasonal subgrade modulus ; E (bit-mon), E (bit-win), E (bit-win) = moduli of subgrade layer during monsoon, winter and summer respectively (MPa).

As can be observed from the above expressions, there is reasonable correlation between the seasonal average subgrade modulus value and the corresponding winter and summer (relatively dry periods) seasonal subgrade modulus value. However, owing to the large variance in the monsoon subgrade modulus value, the correlation between the seasonal mean modulus and monsoon modulus is not good.

The relationships are also presented in Figs. 6 to 8.







5.7. Model for Seasonal Adjustment of Subgrade Modulus
Subgrade modulus usually has the lowest value during monsoon season compared to other seasons. Post monsoon modulus value is commonly used in India for design of pavements and overlays. In order to estimate the monsoon subgrade modulus from other seasonal modulus, the subgrade modulus estimated in different seasons were correlated. Equations 22 and 23 give the relationships for estimating monsoon subgrade modulus from the modulus value obtained in another season.

E(sub_mon) = 28.39+0.2888 E(sub_win) (R2 =0.422) ... Equn. 22
E(sub_mon) = 21.821+0.3641 E(sub_sum) (R2 =0.616) ... Equn. 23
where, E (sub_mon) = subgrade modulus in monsoon (MPa); E (sub_win) = subgrade modulus in winter (MPa); E (sub_sum) = Subgrade modulus in summer (MPa)

6. OVERLAY DESIGN PROCEDURE USING IITKGP FWD
Based on the results obtained in the present study, the following approach is proposed for design of bituminous overlays in India. The mechanistic criteria (fatigue and rutting) adopted in the Indian Roads Congress guidelines2 forms the basis for the proposed method. The following are the steps proposed for design of overlays for Indian highways based on FWD evaluation.
(i.) Measurement surface deflections of the in-service pavement using FWD and to collect pavement layer thicknesses.

(ii.) Normalization the deflections to correspond to standard load of 40 kN.

(iii.) Backcalculation pavement layer moduli from the normalized deflections using backcalculation software (BACKGAIIT).

(iv.) Adjustment of the bituminous layer modulus (backcalculated) to a standard temperature of 35oC using the correction factors recommended in the present study.

(v.) Adjustment of the subgrade modulus to correspond to post-monsoon condition from the model developed in the present study (given by equations 22 and 23).

(vi.) Analysis of the in-service pavement using elastic layer theory with the backcalculated (corrected) moduli and layer thicknesses collected from field as inputs. This includes computation of critical Strains (a) Horizontal Tensile Strain at the bottom fiber of bituminous layer and (b) Vertical Compressive Strain on top of subgrade. The loading configuration and the locations of critical strains considered for analysis are shown in Fig. 9.

(vii.) Estimation of the remaining life of the pavement by using the strain values obtained in step (vi) as inputs in the following performance criteria1.


For design of bituminous overlay, a trial thickness of overlay of an appropriate material (with known modulus value) has to be selected and the critical strains have to be evaluated. Design overlay thickness can be selected by trial in such a way that the critical strains computed with the selected thickness as input will be less than the permissible limits given by the performance criteria for the design traffic level considered. Typical modulus values of bituminous layers obtained in the present research work can be used for analysis. A typical design example is presented in the following paragraphs for better appreciation of the design approach.





7. DESIGN EXAMPLE
(1) Design Traffic : 50 msa
(2) Details of existing pavement : 170 mm of weathered bituminous layer 585 mm of granular base
(3) Condition of the Pavement : Surface cracks (>20 % area)
(4) FWD evaluation conducted in the month of January.
(5) Pavement surface temperature at the time of FWD tests 35oC
(6) Surface deflections measured using FWD and normalized for a load of 40kN



(7) Backcalculated modulus values of in-service pavement layers obtained using BACKGAIIT are given below.




(8) Modulus value of new BC layer for overlay for a temperature of 35oC = 1250 MPa (obtained from the present investigation)
(9) Backcalculated modulus value of the existing bituminous layer = 398 MPa
     (No temperature correction as the existing pavement was cracked)
(10) Backcalculated modulus of the existing Granular Base (corrected for seasonal modulus) = 216 MPa
(11) Backcalculated Subgrade modulus (Adjusted for monsoon condition) = 46 MPa
(12) Selection of overlay thickness:

Since the thickness of the existing granular layer is very large (585mm), rutting will not be a major concern as established in the research findings of Research scheme R-5623 of Ministry of Road Transport and Highways (MORTH). Hence, only fatigue criterion is taken into consideration for selection of overlay thickness.

Trial overlay thicknesses have been selected and maximum tensile strain at the bottom of the overlay has been computed using the thicknesses and moduli of various layers as inputs. The following table shows the trials made with two overlay thicknesses and the tensile strain values computed for the trial thickness selected. Estimated fatigue lives for the trial thicknesses are also given in the table.

(13) Provide overlay of 110 mm thick bituminous layer with layer modulus of 1250 MPa (at 35oC)





8. CONCLUSIONS
This paper deals with a new, rational approach for design of bituminous
overlays for Indian highways. Salient features of an in-vehicle FWD, BACKGAIIT, a program for backcalculation of layer moduli, have been presented in the paper. Details of various investigations conducted on different highways using the in-vehicle FWD have also been presented in the paper. Structural evaluation of in-service pavements using the FWD yielded several useful inputs regarding the moduli values to be assigned to different layers of in-service pavements for different seasons. Valuable information has also been obtained regarding the performance of new technology thick bituminous layers. The steps to be followed for design of bituminous overlays using the FWD evaluation based approach proposed in the present study have been presented. The development of an indigenous FWD system and performance criteria were necessary for the advancement of pavement technology in India and for the adoption of rational overlay design approaches. Easy availability of the low-cost FWD in India is expected to popularize the use of rational tools in the overlay design and evaluation of pavements in India.

ACKNOWLEDGEMENTS
The authors are grateful to the Ministry of Shipping, Road Transport & Highways, Government of India for the financial support provided through research scheme R-81 (Structural Evaluation of Pavements using FWD in Eastern India) for the development of FWD and for collection of field data.

REFERENCES
1. IRC: 81. Tentative Guidelines for Strengthening of Flexible Road Pavements using Benkelman Beam Deflection Technique, Indian Roads Congress, New Delhi, 1997.

2. IRC: 37. Guidelines for Design of Flexible Road Pavements, Indian Roads Congress, New Delhi, India, 2001.

3. Collop, A.C. The Effect of Asphalt Layer Thickness Variation on Pavement Evaluation Using Falling Weight Deflectometer, International Journal of Pavement Engineering, Vol 1(4), 2000, pp.247-263.

4. Stephen F. Brown., Tam, W.S. and Janet M. Brunton. Structural Evaluation and Overlay Design: Analysis and Implement. Proceedings of 6th International Conference on Structural Design of Asphalt Pavements, 1987, pp.1013-1028.

5. Brown, S.F. and Brunton, J.M. An Introduction to the Analytical Design of Bituminous Pavements. 3rd Edition, University of Nottingham, 1996.

6. Richer, C.A. and Irwin, L. Application of Deflection Testing to Overlay Design: A Case Study. Transportation Research Record No. 1196, Transportation Research Board, Washington, DC, 1988, pp. 193-200.

7. Mamlouk, M.S. Overlay Design Method for Flexible Pavements in Arizona. Transportation Research Record No. 1286, Transportation Research Board, Washington, DC, 1990.

8. Sidess, A., Bonjack, H. and Zoltan, G. Overlay Design Procedure for Pavement Maintenance Management Systems. Transportation Research Record No. 1374, Transportation Research Board, 1992, pp.63-70.

9. AASHTO Guide for Design of Pavement Structures, American Association of State Highway and Transportation Officials, 1993.

10. Bayomy F. M., Al-Kandari F. A. and Smith R. M. Mechanistic Overlay Design System for Idaho. Transportation Research Record No. 1543, Transportation Research Board, 1996, pp.10-19.

11. Mahoney, J.P and Pierce, L.M. Examination of State Department of Transportation Transfer Functions for Mechanistic-Empirical Asphalt Concrete Overlay Design. Transportation Research Record No. 1539, Transportation Research Board, Washington, DC, 1996, pp. 25-32.

12. Pierce, L.M. and Mahoney, J.P. Asphalt Concrete Overlay Design Case Studies. Transportation Research Record No. 1543, Transportation Research Board, Washington, DC, 1996, pp. 3-9

13. Corley-Lay J. B. Efforts by North Carolina Department of Transportation to Develop Mechanistic Pavement Design System. Transportation Research Record No. 1539, Transportation Research Board, 1996, pp.18-24.

14. Arnold G. Design of Rehabilitation Treatments for New Zealand’s Thin-Surfaced Unbound Granular Pavements. Transportation Research Record No. 1652, Vol. 2, Transportation Research Board, 1999, pp.43-50.

15. Abdallah, I., Ferregut, C., Nazarian, S. and Lucero, O. M.. Prediction of Remaining Life of Flexible Pavements with Artificial Neural Networks Models. Nondestructive Testing of Pavements and Backcalculation of Moduli, 3 rd Volume, Eds. Tayabji, S.D. and Lukanen, E. O, ASTM STP 1375, ASTM, West Conshohocken, PA, 2000, pp. 484-498.

16. Amaranatha Reddy, M. Srinivasa Kumar, R., Sudhakar Reddy, K. and Pandey, B.B. A Low Cost Falling Weight Deflectometer for Pavement Evaluation in India, International Journal of Pavement Engineering & Asphalt Technology, UK, Vol. 4, issue 1, May 2003, pp.44-54.

17. Goldberg, D. E. Genetic Algorithms in Search, Optimization and Machine Learning. Addison-Wesley, International Student Edition, 2000.

18. M. Amaranatha Reddy, K. Sudhakar Reddy and B.B. Pandey. Evaluatio of Effective Pavement Layer Moduli Using Genetic Algorithms, International Journal of Pavement Engineering & Asphalt Technology, UK, Vol. 3, Issue 1, October 2002, pp. 6-19.

19. Reddy, K.S. Analytical Evaluation of Flexible Pavements Ph.D. Thesis (1993), Indian Institute of Technology, Kharagpur, India, 1993.

20. Reddy, M.A., Reddy, K.S. and Pandey B.B. Selection of Genetic Algorithm Parameters for Backcalculation of Pavement Layer Moduli, International Journal of Pavement Engineering, Vol. 5, June 2004, pp. 81-90.

21. M. Amaranatha Reddy, K. Sudhakar Reddy, and B.B. Pandey “Genetic Algorithms for Backcalculated moduli of Pavement Layer Moduli”, Highway Research Bulletin, Indian Roads Congress, Vol. 66, pp1-10, 2002.

22. Specifications for Road and Bridge Works, 4th Edition, Indian Roads Congress, 2001.

23. Final report of the Road Research Scheme R-56, Development of a Computer Program and Design Charts for Analytical Design of Flexible Pavement. Submitted to the Ministry of Surface Transport by IIT, Kharagpur, 1999.