**6. Uncertainty in cannery wastewater**

There are several types of uncertainties that should be addressed during the design of a wastewater treatment plant, e.g. the variation in strength and quantity of wastewater entering into the plant, the diversity, and the dynamics of the microbial community. An uncertainty analysis for a pre-denitrification plant that uses an activated sludge unit was performed [18]. The unit consists of five compartments: the first two are anoxic and the last three are aerobic. Three scenarios were considered in that study that cover the uncertainty due to stoichiometric, bio-kinetic and influent parameters; uncertainty due to hydraulic behavior of the plant and mass transfer parameters; and uncertainty due to the combination of both scenarios. The study concluded that parameters related to the first and second scenarios introduce significant uncertainties in the plant performance measures. In addition, it was stated that the applied uncertainty farming technique largely affects the uncertainty estimates.

The Monte Carlo simulation was intensively used to simulate the design and upgrade of wastewater treatment plants under uncertainty in balancing effluent costs, violating effluent quality standards, predicting the disinfection performance, generating different influent compositions for posterior process performance evaluation or as a pragmatic procedure to automate the calibration of ASM models, and considering the impact of the input parameter uncertainty on the multi-criteria evaluation of control strategies at wastewater plant [19–22].

Due to the complexity and non-linearity of wastewater treatment plant operations, mathematical models are generally not sufficient to predict the performance of WWTPs. Therefore, AI models have been proposed as an alternative model to linear methods. The methods for minimizing the effect of uncertainty in wastewater characteristics and wastewater flow on wastewater treatment reported in the literature have included support vector machine (SVM) and artificial neural network (ANN) [23]. An optimization model to control uncertainty in operation of wastewater treatment from the shale gas production has been reported in the literature [24]. In addition, genetic algorithms have been developed to model and optimize a biological wastewater treatment plant [25].

Information regarding uncertainty in the wastewater treatment plants in treating cannery wastewater is lacking in the literature. However, the principles governing uncertainty in wastewater treatment plants can be applied to control uncertainty in cannery wastewater treatment.
