The deflection response of pavements under an applied load will be studied in the strategic highway research program's (shrp's) long-term pavement performance (ltpp) study using a falling weight deflectometer (fwd). The shrp computer software system for the collection of data also possesses the capability to filterthe data by means of a digital low-pass filter. Shrp decided to assess the effects of digital filtering of fwd data before implementation of the software in the field. Comparative analyses of noise were performed using the results of an fwd pilot study conducted in greensboro, north carolina, as well as other test sections throughout the united states. The results show that filtering of fwd data introduces significant random errors, particularly for rigid pavements tested under heavy loads. These errors tend to compound each other when the filtered deflections are normalized by the filtered load data. Also, the use of filtered loadand deflection data may yield normalized deflection responses that exceed current normalized deflection tolerance limits, particularly for heavy loads on rigidpavements. Accordingly, it has been recommended that all fwd data be collected by shrp with the filter off. However, because some unknown level of noise is contained within fwd data, it is also recommended that additional load- and deflection-time histories be collected and stored. Thus, if advances occur in the filtering process, the data can be reanalyzed to obtain more accurate peak load and deflection values for use in the backcalculation of layer moduli. This paper appears in transportation research record no. 1260, Measurement of pavement surface condition 1990.
Abstract