AN IMPROVED METHOD OF CULWAY CALIBRATION . 15TH ARRB CONFERENCE, DARWIN, NORTHERN TERRITORY, 26-31 AUGUST 1990; PROCEEDINGS PARTS 1 TO 7

The CULWAY Weigh In Motion (WIM) system was originally conceived and tested in Western Australia by the Main Roads Department (Peters 1986) and further developed by the Australian Road Research Board. The system has been installed by all State Road Authorities and is producing data for numerous purposes. The CULWAY system uses a set of Mechanical Strain Amplifiers (MSA's), affixed to the soffits of culvert units beneath a road, and axle sensors to weigh and classify vehicles. The system has no means of calculating the transverse vehicle position. The accuracy of a WIM system is difficult to assess, particularly when vehicle transverse position variation can have a significant effect on the calculated mass of the vehicle. This paper proposes that before an accuracy can be quoted, it is important to be able to 'weigh' the same vehicle with a high degree of repeatability. The degree of confidence that can be placed on the data should then specify the allowable positions of the vehicles for which the quoted accuracy applies. During the calibration of CULWAY sites it was found that the same vehicle travelling at the same speed, in different transverse positions on the roadway, would produce a variation of strain readings, accumulated from all MSA's, of up to 30 per cent. It was found that in a group of instrumented culvert units, it was common for none of the gauges to produce peak readings of similar magnitude even though all gauges were set to the same output value under a constant applied strain. Strain responses on each gauge were plotted against tyre position in order to establish a relationship. It was found that a logarithmic polynomial curve fitted the initial test data well. The recorded data at all of seven CULWAY sites in Western Australia was assessed and found to conform well to the type of curve chosen. The generated curves were then used to optimise the reading uniformity, within the range of most probable vehicle positions, without requiring the vehicle position information. The improvements in the CULWAY calibration procedure as outlined in this paper have resulted in some dramatic improvements in the accuracy of some CULWAY sites. The variation in measured vehicle mass, over a one metre transverse position difference, has been reduced from as much as 30 per cent to a maximum of 10 per cent. Shallow soil cover sites appear to benefit most from the techniques used as they tend to exhibit greatest initial transverse variability. (Author/TRRL)