a. Incorrect or inconsistent b. Height differential between rail gauge. rails. 

Context in source publication

Context 1

... surveys have been carried out by various means and for a variety of applications for decades. Although optical offset and photogrammetric methods are successfully used in many cases, triangulation using theodolites or total stations is preferred for most alignment surveys carried out by surveyors. This paper describes a novel alignment survey which required 300m long crane rails to be measured to an accuracy of one millimetre using a combination of angular triangulation and distance measurement. In late 1992 the Department of Geomatics was approached by a major Australian manufacturer with a somewhat unusual problem. The problem concerned the excessive and uneven wearing of the wheels of an overhead gantry crane in a main storage and dispatch warehouse. The replacement of the overhead crane wheels is both an expensive and time consuming process, which of course also results in a significant loss of production time. It was suggested that two primary factors were responsible for the excessive and uneven wear of the crane wheels. First, departures from the standard rail gauge (Figure 1a) would result in wear due to excessive lateral force. Second, height differentials between the rails (Figure 1b) would result in wear due to uneven distribution of weight. used by engineers on site to evaluate and correct any irregularities present at each of the rail positions. To formulate a suitable measurement strategy the first step was to inspect the working environment. The crane rails are approximately 300m long and are mounted about 7m above the ground on steel girders. Supporting uprights are spaced at 9.1m intervals and the separation between the rails is approximately 40m, slightly less than the width of the warehouse. From a site inspection it is possible to identify and minimise any potential problems, both large and small, during the survey. In this instance it was readily apparent that the floor area was far too cluttered with manufactured product to permit the establishment of instrument stations with unobstructed lines of sight. To counteract this problem it was negotiated that a five to ten metre buffer zone be established along the length of the warehouse prior to commencing the job. In ideal circumstances the warehouse would have been substantially cleared of product, but this was simply not commercially feasible for an area of 1.2 hectares. Access to the crane rails was to be via two 8 metre ladders to be provided by the client. Although this was likely to be awkward given the amount of product on the warehouse floor, there was no immediate alternative. Possible sites for permanent reference targets were identified. Some were accessible from floor level or using the ladders, but a number of high targets would have to be mounted to ensure clear lines of sight. It was identified on site that some targets would likely require use of the crane platform for a short time prior to the day of the survey. Also during the inspection it was agreed that the two crane platforms, which would not be operating during the survey, would be parked at the ends of the warehouse to minimise line of sight obstructions. A consequence of this would be that the extreme ends of the rails would be inaccessible, reducing the overall length to 270 metres. Some preliminary ideas for the design of the survey were discussed on site but no immediate decisions were made. Site plans were supplied by the company to facilitate a more detailed consideration of the design of the survey. Having inspected the site it was necessary to establish a measurement methodology capable of satisfying the accuracy and organisational requirements of the task. The rail position and length obviously raised a number geometric problems that needed to be addressed. Initially the use of an on-line triangulation system, comprising three total stations and PC based software, was considered. The principles of such systems are very well known (Allan, 1988) and they have gained widespread acceptance for industrial measurement applications (Roberts and Moffitt, 1987; Woodward, 1987). Despite of the elongated geometry, such a system could quite easily satisfy the accuracy requirements and coordinate determination could be provided in real time. The advantage of the online system is that it provides a quality measure of every intersection measurement as it is taken, and measurements can be repeated as needed. However the online system was deemed unsuitable because it could not satisfy many of the other job requirements. The amount of time needed to establish the system was the primary concern. The establishment of a three instrument system set up requires approximately forty-five minutes and it was estimated that approximately twelve set ups would be necessary to adequately cover the length of the rails. In addition, without additional measurements at each set up the potential for propagation of error from the transfer of a local coordinates between set ups was an issue. The cost associated with the increased time at the site made the use of this technique uncompetitive. A variation on the online triangulation system was considered as a means of bypassing some of the problems associated with the method. The proposed method would still be instrument based, however each instrument station would be established pseudo- independently. This variation was based primarily on the establishment of multiple instrument stations through resection to a number of common points, followed by intersection observations to the rails. Eight resection targets strategically distributed throughout the warehouse were considered sufficient to ensure that every instrument station would have at least six visible reference points (see Figure 2). Information gathered at each instrument would then be stored in a data recorder. This information could be downloaded and manipulated in a survey network adjustment to obtain station coordinates for each point. Similar approaches have been successfully applied to industrial measurement tasks elsewhere (Shortis, 1992; Stirling ...