| OCR Text |
Show The model is constructed largely of perspex (plexiglas) to the actual manufacturing drawings of the full scale plants, verified by site visits as required. It is not unknown for site erection staff to re-design a duct system, unknown to the original engineers and drawing office. It is absolutely essential that the model is totally accurate internally to the full size plant and that any internal buckstays, stiffeners, ladders, risers, downcomers etc, are faithfully reproduced. In the early days of model testing a considerable number of models had to be rebuilt and re-tested following last minute site revelations. Once the model is constructed it is connected to a suitable calibrated air supply. The model is then flowed on air utilising equivalent air flow velocities and pressure losses as the full size plant Fig. 2. By the use of miniature pitot tubes Fig. 4 and electronic micromanometers careful velocity mapping of the air flow from the burner exits is carried out. Typically 24 points on 3 radii are employed. At the same time, using in-house developed instruments, the degree of swirl present in the burner exits is evaluated and unstable flow conditions isolated. Depending on the type of model and the complexity, velocity mapping is also carried out at other points of interest i.e. duct bifurcations, abrupt changes of section etc. Smoke injection may also be utilised coupled with high speed photography and video recording if areas of severe problems are anticipated. The results from the above are then correlated and plotted via a simple computer programme. The resulting polar diagrams Fig. 6 I and 6 F give a very effective representation of the flow regimes in the burners and the base line discrepancy of flows between burners. Flow discrepancies of ± 20% are not unusual between adjacent burners even when mounted in a common windbox and far higher variations have been measured at times. - 7 - |
