Coal Preparation » Process Control
It has been recognised for many years that the performance of 'sputnik' type hydraulic raw coal distributors is extremely poor. A number of modifications have been made to the basic design at different sites. Operators also try to improve performance by making adjustments to inlet water flow rates, and at some sites to outlet breather pipe valves, while monitoring desliming screens for evidence of poor distribution (bias). Recognising that these operational practices were widespread, but unproven, ACARP Project C4055 entitled 'Control of Hydraulic Distributors to Minimise Feed Bias' commenced in July 1995. The stated objectives in the original proposal to ACARP were:
- to determine if desliming screen bed depth can be used to infer the flow rate of coal over the screen, and hence distribution bias,
- to investigate if breather pipe manual valves or tangential inlet water flow rates can be used as manipulated variables to control distribution,
- to devise a suitable control strategy, assuming the first two objectives were achieved.
Project test work was carried out at Oaky Creek coal preparation plant and in the JKMRC pilot plant.
During the site test work at Oaky Creek, large scale sampling techniques were devised to measure accurately both the water and solids flow rates out of the distributor. It was established that a significant bias occurred in the water flow rate to the different modules of the plant, a result which had not been anticipated. A Milltronics Multiranger Plus ultrasonic level measuring instrument was evaluated as means of measuring desliming screen bed depth, and depth versus oversize mass flow rate calibrations were established for each of the six desliming screens. The results showed that it is possible to monitor distribution bias on-line using readily available and inexpensive instrumentation, thereby achieving the first project objective.
A 1/5 scale model of the Oaky Creek distributor was constructed in clear perspex at the JKMRC, and evaluated in terms of residence time distribution, mixing and solid and water distribution by comparison with Oaky Creek data.
The model satisfactorily represented the performance of the full scale unit.
Test work using the model clearly showed that altering inlet water flow rates had no effect on distributor bias. The model test work also showed that while breather pipes were essential for best performance, their use as a manipulated variable for control purposes was unlikely to be successful. Thus the second objective of the project was not achieved, and hence the third objective was not pursued further.
The project work, involving the use of a model which allowed visualisation of the internal flow, showed that the design of the sputnik raw coal distributor is fundamentally flawed. Modifications may alleviate the poor performance to a certain extent, however consideration should be given to developing and proving an alternative design.