Resources

This technical advice note details the methodology to convert the traffic speed deflectometer (TSD) data to the more familiar falling weight deflectometer (FWD) output, enabling any TSD data collected to be directly compared to any historically collected FWD dataset.

Background

The traffic speed deflectometer has been used for three years on New Zealand roads to generate a deflection bowl similar to the falling weight deflectometer bowls.

However, owing to the differences between the FWD and TSD machines in terms of the load configuration, the types of sensors used to determine deflection bowls, and the differences in pavement response between the static and dynamic testing method, a conversion is required to generate a comparable output. The TSD device records the deflection slopes measured from the device’s doppler lasers, and a deflection bowl is calculated using integration, whereas the FWD vertical deflections are measured by geophones.

The TSD loads recorded in RAMM are not adjusted to a standardised 40 kN load whereas the FWD results are typically normalised. Therefore, environmental factors such as forces induced by wind, road roughness, cornering and vehicle accelerations can cause changes in the axle loads. The TSD bowl should be adjusted to a standard 40 kN load prior to any analysis being performed. This is simply performed by multiplying the deflections, at each of the offsets, by the factor obtained by dividing 40 kN by the estimated load, obtained from RAMM, for that particular bowl.

The dataset received from the supplier contains two different deflection bowls:

• the Greenwood bowl which considers any asymmetry of the bowl and extends from the bowl centre to 900mm offset. However, no deflections are calculated for the customary FWD offsets at 1200 and 1500mm
• the ARRB bowl which assumes all bowls are symmetrical but does determine deflections at both 1200 and 1500mm offsets

Users of the data, therefore, have the option of using somewhat different bowls. Because the technology is new, there is not yet a consensus on which interpretation provides the more appropriate bowl for any specific circumstance, so it will be important for users to note which option they are adopting for any project.

Early trials with New Zealand data indicated that each of the two bowls could give more typical FWD bowl shapes in different circumstances, but on balance the ARRB bowl provided slightly more consistency. For that reason, only the ARRB bowl was used in subsequent transformations to 'equivalent FWD' bowls.

The equivalent bowl may be used by practitioners with back-calculation software, layer modular ratios and fatigue criteria in the same way as if the data were obtained from traditional FWD equipment which has been used to populate the RAMM database over the last 20 years. However, the bowls should be used with caution until New Zealand gains more familiarity with the results from the traffic speed deflectometer results. This is particularly true when calculating remaining lives of pavements; in this situation the estimated remaining lives should be used as a ranking tool rather than absolute value.

The deflection data contained in RAMM is the ARRB generated bowl.

Reliability ranking of bowl shapes

The reliability of any given bowl may be judged by inspection of the goodness of fit, which quantifies the relative compliance of any measured bowl with any other bowl generated from a methodical approach (eg integration of deflection slopes, forward analysis of moduli/layer thickness system etc).

Typical ranges for NZ TSD data are:

Table 1: goodness of fit – typical ranges

This parameter is useful for determining when decisions should give more weight to factors other than the TSD information, ie the visual survey and experience with historic performance, as well as indicating the need for as-built information or additional subsurface investigations.

When comparing TSD data on NZ state highways in successive years (2015 and 2016), the general trend is reasonably strong.

However, as can be clearly seen in the following example, both TSD sets seem to be generally underestimating the FWD deflection (as expected, given differences between the contact area configurations of the two devices), but the TSD distributions also highlight differences between successive years.

Deflection bowl conversion

Several methods for transforming New Zealand TSD deflection bowls, d(r), into equivalent FWD deflection bowls have been examined and the network specific individual offset method is the preferred method. It provides a strong transformation from the ARRB calculated deflection bowl to an equivalent FWD deflection for the given offsets.

Composite surface modulus values are used in the conversion and are calculated using equations 1.1 and 1.2.

1.1
1.2

Where: 

E₀(r) is the surface modulus at a distance of r (in mm) from the centre of the loading plate.

µ is Poisson’s ratio (set equal to 0.35)

σ₀ is the contact stress in MPa (set to 0.5659 MPa)

a is the radius (in mm) of the loading plate (set to 150 mm), and

d(r) is the deflection at the distance r (mm).

The k(r) values, displayed in table 2, are used in the transformation of the deflections generated using the ARRB methodology from the traffic speed deflectometer into bowls equivalent to those generated by a falling weight deflectometer. To generate FWD equivalent bowls from the TSD then the TSD deflections must be standardised to 40 kN load first. As mentioned above this is performed by multiplying the deflections, at each of the offsets, by the factor obtained by dividing 40 kN by the estimated load, obtained from RAMM, for that particular bowl.

Table 2: input k factors

Equivalent weight falling deflectometer bowls

Equivalent falling weight deflectometer bowls can be generated from the TSD deflection bowls using equations 2.1 through to 2.9 to generate the deflection d’(r) at the distance r (in mm) from the centre of the load.

Note: the constants used in the following equations are calibrated to a subset of the national data base, and the resultant deflection bowl shape should be scrutinised carefully.

2.1
2.2
2.3
2.4
2.5
2.6
2.7
2.8
2.9

Where d’(r) are the FWD equivalent bowls and d(r) are the ARRB derived TSD bowl values at the distance r (mm) from the centre of the load. The k values are found in table 2 above.

Use of TSD deflection data for pavement analysis

While the equivalent FWD bowls can be used in pavement analysis, the results, at this stage, should be treated as a ranking tool rather than providing absolute results. Furthermore, year by year comparative analysis is not currently viable due to some differences between the 2014/15 and 2015/16 results that are unresolved at this stage.

Implementation of the transformation methodology can be performed using the associated spreadsheet: Traffic Speed Deflectometer – NZ Interpretation [XLSX, 34 KB].

Additional information on the development of the TSD conversion can be found in the following document: Traffic Speed Deflectometer - NZ Interpretation Final [PDF, 9.7 MB]. 

Further information

For further information contact Martin Gribble, Principal Pavement Engineer, or call  04 4894 6256. 

John Donbavand