Strategy to identify the causes and to solve a sludge granulation problem in methanogenic reactors: application to a full-scale plant treating cheese wastewater.

Granulation of biomass is at the basis of the operation of the most successful anaerobic systems (UASB, EGSB and IC reactors) applied worldwide for wastewater treatment. Despite of decades of studies of the biomass granulation process, it is still not fully understood and controlled. "Degranulation/lack of granulation" is a problem that occurs sometimes in anaerobic systems resulting often in heavy loss of biomass and poor treatment efficiencies or even complete reactor failure. Such a problem occurred in Mexico in two full-scale UASB reactors treating cheese wastewater. A close follow-up of the plant was performed to try to identify the factors responsible for the phenomenon. Basically, the list of possible causes to a granulation problem that were investigated can be classified amongst nutritional, i.e. related to wastewater composition (e.g. deficiency or excess of macronutrients or micronutrients, too high COD proportion due to proteins or volatile fatty acids, high ammonium, sulphate or fat concentrations), operational (excessive loading rate, sub- or over-optimal water upflow velocity) and structural (poor hydraulic design of the plant). Despite of an intensive search, the causes of the granulation problems could not be identified. The present case remains however an example of the strategy that must be followed to identify these causes and could be used as a guide for plant operators or consultants who are confronted with a similar situation independently of the type of wastewater. According to a large literature based on successful experiments at lab scale, an attempt to artificially granulate the industrial reactor biomass through the dosage of a cationic polymer was also tested but equally failed. Instead of promoting granulation, the dosage caused a heavy sludge flotation. This shows that the scaling of such a procedure from lab to real scale cannot be advised right away unless its operability at such a scale can be demonstrated.