Filamentous bulking caused by Thiothrix species is efficiently controlled in full-scale wastewater treatment plants by implementing a sludge densification strategy

Abstract

Filamentous bulking caused by Thiothrix species is responsible for sludge washout and loss of performance in dairy wastewater treatment plants. A long-term study was conducted over 1.5 years to test three different mitigation strategies in a full-scale plant composed of two parallel sequential batch reactors (SBR1 and 2). Strategies based on polyaluminium chloride addition and volatile fatty acids reduction were ineffective to permanently solve the problem. On the contrary, modification of the reactor cycle based on the implementation of a periodic starvation proved efficient to solve the biomass wash-out and drastically reduce the sludge volume index in both reactors. Bacterial diversity analysis using 16S amplicon sequencing and quantitative PCR indicated a reduction of Thiothrix abundance from 51.9 to 1.0% in SBR1 and from 71.8 to 0.6% in SBR2. Simultaneously, the abundance of the glycogen-accumulating bacterium Candidatus Competibacter increased in both reactors. Microscopy analysis confirmed the transition between a bulking sludge towards a granular-like sludge. This study confirms the applicability of a periodic starvation to (1) solve recurring Thiothrix bulking, (2) convert loose aggregates into dense and compact granular-like structures and (3) considerably reduce energy demand for aeration.

Figure 1

Flow diagram of the dairy wastewater treatment plant. The raw wastewater was screened then collected in an equalization buffer tank. During production, a second buffer tank received a continuous flow from the dissolved air flotation (DAF) unit. The wastewater was then treated with two conventional sequential batch reactors (SBR1 and SBR2) operated in parallel. The effluent was collected in a buffer tank, screened through a drum filter and released in the receiving watercourse.

Figure 2

Sludge volume index (SVI, •) profile and TSS in mixed liquor (grey lines) of SBR1 (top) and SBR2 (bottom) over the experimental period. Mitigation strategies consisted of the addition of polyaluminium chloride (I), inhibition of VFA production (II) and modification of the feeding pattern with (IIIa) or without (IIIb) aeration in the buffer tanks. SVI values were measured after 30 min settling. Day 1 corresponds to January 1^st, 2015.

Figure 3

Distribution of the bacterial phyla in SBR1 (top) and SBR2 (bottom) between days 278 and 484. The abundance of each phylum is expressed as the relative gene frequency of the corresponding OTUs.

Figure 4

Temporal change in the relative gene frequency in SBR1 (left) and SBR2 (right) of Skermania spp. and Thiothrix spp. (A and B), Defluviicoccus spp. and Candidatus Competibacter (C and D), Tetrasphaera spp. and Candidatus Accumulibacter (E and F).

Figure 5

Temporal change in the concentration of Thiothrix eikelboomii (A) and Candidatus Accumulibacter (B) expressed in number of copies of 16S rRNA operon per mg TSS, in SBR1 and SBR2 from day 250 to 500. Densification strategy (Strategy III) was applied from day 283 on SBR2 and from day 358 on SBR1.