No construction is subject to completely static forces, even buildings are subject to variations in wind load, live weight, seismic tremors, vehicle movements, differential heating and cooling, humidity changes and even algal growth cycle
Every force applied to a material (stress) results in some accommodation in or deformation of that material (strain). Stress produces strain. If the strain is excessive, small cracks are produced. With repeated stress - strain cycles - these cracks will grow. Over a given period, repeated small stress cycles may cause significant crack accumulation, whereas larger stress cycles will lead to significant plastic deformation, work hardening and failure. Where the material is metal, this is called metal fatigue and has been the subject of much investigation, which has shown that the number and range of cycles (cycle density profile) is much more relevant than cycle rate in producing failure.
While buildings are subjected to a relatively minimal set of dynamic forces, there are few constructions that are subject to a more varied set of forces than ships: engine, shaft and propeller vibrations, wave impact and slams, swell length and amplitude, wind loading, cargo and passenger loading and movement, diurnal temperature variation, and longer term; temperature and salinity variation of water. Swell in particular, of the order of the ship length and of sufficient amplitude, can produce massive stress cycles on a ship, alternately supporting the hull at bow and stern, and midships. This can lead to rapid hull or superstructure failure. There is a celebrated case in Australia's early post-World War II Antarctic activities, where the tired military surplus landing craft was returning North through days of heavy swell. The Chief Engineer (and any person aboard with one eye and half an ear, it can be imagined) noticed transverse cracks, each several feet long, in the superstructure cycling open and closed with passage through the swell. They were still days from Fremantle, the nearest port and the ship was likely to sink within hours if nothing was done. The cycling could not be stopped but the material could be reinforced. Thick plates were manhandled into place over the ever-growing cracks, holes marked and drilled and with one side bolted, they waited for the crack to close again. At the point of squealing shut, all bolts on the mobile side were slammed home from above to those anxiously waiting below with nuts and spanners. The nuts were tightened before the next opening cycle and the ship limped home with all hands.
It is no wonder that ship builders (and their passengers) want to avoid incidents of the kind related above. In particular, they want to balance the cost of producing a ship with its projected life in a given maritime environment and type of service. For example, a ship's service life may be considerably longer in the quieter waters of the Mediterranean, than in the harsher Atlantic. This is not because the material that the ship is made of differs (usually steel) between environments but the stress cycle history or cycle density accumulates more rapidly in the more stressful environment. Metals of different types, alloys, process and geometry can be subjected to measured stress cycles which leads to the development of a cycle density profile beyond which the metal is likely to fail.
These theoretical profiles are well known for a number of metals including steel and aluminium. What is less well known, or at least more difficult or expensive to measure, until recently, is the actual in-service accumulation of stress cycles. By measuring the stress cycle that the ship is exposed to in realtime, the ship operator can detect structural weakness before it becomes failure, extend service life without fear of fatigue failure, or see the effect of changed service use or environment. With such information a ship designer may be able to improve a design, substituting materials, strengthening stressed points, removing structural weight where over-designed, increasing cargo mass and distribution, or varying engine power. But until very recently there have only been solutions available that cost tens of thousands of dollars for realtime cycle analysis, or less expensive and far less convenient post-processing solutions that must deal with immense amounts of stored and transferred data.
In Australia, dataTaker became aware that Incat Tasmania P/L was looking for a cost-effective solution that would allow them to perform cycle counting on each of the ships that they produced. Incat is a world leader in aluminium wave-piercing catamaran technology, and the largest employer in Tasmania's capital city.
Based around a Motorola PowerPC running at 48 MHz, with up to 42 analog acquisition channels, megabytes of PC card DOS file storage, and flexible remote communications capability, the DT800 was an ideal balance of power and capability, coming in at about one seventh the cost of the dedicated system that had been previously trialled. In close cooperation with Incat's R&D department, the engineering team at dataTaker's head office in Melbourne, Australia, implemented a realtime form of the widely recognised Rainflow cycle counting algorithm. Implemented in accordance with International Standard ASTM E 1049 'Standard practices for cycle counting in fatigue analysis', the DT800 is able to support eight to 10 strain gauges sampling at 20 Hz in realtime.
According to John Bradfield, dataTaker's Chief Engineer, "Even though we have been building world-class data loggers since 1983, we have not had the computing platform until now to do the necessary calculations in realtime. The DT800 allows the reduction of gathered stress data into a desired cycle histogram - megabytes of raw strain gauge data can be reduced to a few tens of cycle count values - you can imagine what this means in terms of minimising storage space and data transmission time. Likewise, post-processing becomes non-existent with the issuing of a single command producing a simple profile report. One of the most significant features that we built into the cycle counting code was the user's ability to reject cycles with sizes below a specified threshold. When that approach was taken in conjunction with a custom peak and valley detection algorithm, the points presented to the Rainflow analysis code was minimal."
One of the further advantages of the dataTaker approach to Rainflow analysis is that the analysis code works in engineering units familiar to the user and is completely divorced from any particular sensor implementation. This means that the analysis can be applied to all types and wiring of bridges, extensometers, clinometers, or even the results of realtime calculations involving the fusion of a number of sensors.
dataTaker's Director of Engineering, Graham Henstridge, is enthusiastic about the possibilities, "The DT800 is presently limited to between 160 to 200 acquisitions and analysis calculations per second across all Rainflow enabled channels, but we see this increasing to between 500 and 1000 analyses in the next 12 months. This means that while the DT800 is presently suitable for relatively long period cycle counting on large structures like ships, steel bridges and high-rise buildings, in a few months we will be able to consider projects with much shorter cycle periods."
© Technews Publishing (Pty) Ltd | All Rights Reserved
printer friendly version