Filters Italian Company Testori Has Tested The Effects of Weave and Yarn Type in Kelly Filters
Filtration is a key unit operation in alumina plants. This article shows that choosing the appropriate filter fabric has a lasting effect on the filtration performance.
In the dominant Bayer process, for instance, hot caustic soda is added to crushed bauxite to produce a solution of sodium aluminate plus an insoluble fraction known as “red mud”. Thickeners and filters then separate the aluminate from the impurities, which typically make up the greater part of the original bauxite. To complete the process, the aluminate solution is cooled to precipitate aluminum hydroxide, which is then calcined to yield pure alumina (aluminum oxide).
The overflow from the thickeners is typically polished using Kelly filters: horizontal leaf-type filters in pressure-tight housings, with a track which allows the leaves to be pulled out for cake discharge at the end of each batch.
To improve performance when using a mixture of bauxite types from different sources, Testori was asked by the operator of an alumina plant in Portovesme, Sardinia to find an alternative to the standard fabric used in Kelly filters. The aim was to decrease the amount of solids in the filtrate while at the same time reducing blinding of the fabric and extending the operating time between cleaning cycles.
Testori’s standard product for Kelly filters in alumina plants is a polypropylene fabric made from multifilament yarn in a twill weave, with a weight of 310 g/m2 and an air permeability of 130 l/dm2·min. Possible alternatives could be based on:
- different yarn type: spun yarns improve solids retention;
- different weave: satin weave improves cake release, whereas plain weave improves solids retention;
- different permeability.
Testing Seven Alternative Fabrics
With this in mind Testori tested seven alternative fabrics, using a Buchner funnel loaded with 50 ml of thickener overflow and filter aid from the plant, and kept at a temperature of 80 °C. The filter specialists recorded the filtration time and measured the solids concentration using an atomic absorption spectrophotometer. To compare the results with those from the standard fabric, an efficiency parameter E was calculated:
E = (cref/c)/(Δt/Δtref), where c is concentration and t is time.
After the initial tests three of the new fabrics quickly were eliminated. Many further tests narrowed the choice down to two fabrics: A and B. Eventually Testori found that fabric B was the winner in terms of the efficiency parameter E. This fabric is a plain weave made from multistrand yarn. It is lighter than the standard fabric, and although it has a much lower permeability, the filtration time is close to that of the standard fabric.
Accordingly, Testori supplied the company with a complete set of filter cloths made from fabric B. Over the coming months the Italian filter specialists will be testing these to check that the laboratory results are repeated at production scale.