Historically it has been shown on many projects that centrifuges can outperform belt presses in dewatering digested sludge. However results from a number of projects where sludge has been treated before digestion with Cambi thermal hydrolysis process (THP) have shown good results for belt press dewatering – generally greater than 30% DS. The authors propose that the main mechanism in THP is in changing the structure of extra cellular polymer (ECP) that binds water mainly to secondary sludge and limits dewaterability and compressibility of the mixed sludge. The digested sludge after THP also has much finerin structure and this can lead to problems with capture rate, higher polymer rate with some centrifuges. The aim of this work was to aid process selection for final dewatering for future projects to ensure a cake of greater that 30% DS and with a good capture rate.
At Cotton Valley WWTP in the UK, Cambi THP was retrofitted to an existing digestion plant in order to increase throughput from less than 10,000 tds per year to over 20,000 tds by importation of other sludges and to make a class A product. Existing centrifuge dewatering was retained as the plant had to keep producing digested cake during the retrofit. One new centrifuge was added. The site had poor dewatering performance historically – average 22% DS and it was expected that the dewaterability would increase to >30% DS based on experience retrofitting Cambi in other projects where 10% points increase was typical. However to date the plant has averaged 26.7% DS and has not been able to achieve greater than 29% DS even with the new centrifuge and has problems achieving good capture rate. The site has a polymer limitaion of 12lbs/tds which does not help the situation, but increasing polymer seems to lead to unstable results.
Trials of other dewatering technologies were arranged to ascertain whether the problem was sludge related or equipment related. Full scale trials of a commercially available belt press and a commercially available centrifuge were carried out in the fall of 2008 and the spring of 2009.
A 1 meter high performance belt press was tested for a one week period at Cotton Valley in November 2008. The belt press gave 28-32% with the onsite polymer and up to 36% DS with a new polymer. Cambi digested sludge is at 6% so that at normal belt feed rates the loading on the 1 meter belt was up to 1,700 lbs dry solids/h/meter belt. By reducing the load rate to 1050 lbs/h/meter belt the very high dewaterability was achieved. Average polymer dose was about 6 lbs/tds. Observed filtrate quality was higher than the existing centrifuge operation. The cake was visibly of much higher quality and would store more easily and be more useful as a soil improver.
A two week trial was carried out in June 2009 of a medium size rental centrifuge. Some preliminary flocculation trials were carried out that suggested that the sludge was difficult to floc completely and that the flocculent dose would be high. Based on this, the selection of the centrifuge was made in order to give as much possibility of flocculent dilution and flexibility of injection points. In addition trials were carried out with two flocculent vendors to compare a range of polymers including the on site polymer.
The trials showed that by using a large dilution of polymer (0.1% concentration) and splitting polymer dose between pre-flocculation and in bowl flocculation it was possible to achieve cakes in the range of 30-32% consistently and up to 34% by increasing torque. Similar to the belt press trial the vendor was able to demonstrate higher than normal feed rate to a centrifuge that was normally solids constrained at 1.5 tds/tds - the actual rate was demonstrated up to 1.7 tds/h. Polymer doses were used in the range of 12-18 lbs/tds with a dose of 16.6 lbs/tds being optimal for gel polymer. This could be reduced to 13.4 lbs/tds using a bead type polymer. It was possible to achieve above 98% capture consistently.