Exploring the circularity of brine valorisation through neural network modelling of an electrodialysis with bipolar membranes pilot plant

Electrodialysis with bipolar membranes (EDBM) is an innovative electro-membrane technology capable of simultaneously producing acid and base streams. EDBM can be employed for in situ production of chemicals, reducing transportation and storage costs, or integrated with other technologies into circular approaches for the valorisation of brines, recovering high-value materials. Innovative schemes, which foresee feeding the chemicals compartment with a brine instead of water, can be adopted in those cases in which salty acid and/or alkaline solutions can be utilised. In this way, the brine treatment capacity of the process is increased and, at the same time, the water consumption is reduced. The aim of the present work is to model EDBM's behaviour when these innovative saline feed schemes are employed. Due to the challenge of describing the process with a mechanistic approach, related to the non-ideal behaviour of the membranes and the ion interactions in multi-ionic systems, a data-driven model, based on neural networks, was developed. A set of data obtained from a semi-industrial scale EDBM unit was utilised to train and validate a neural network model. The network is capable of describing the different behaviours with high accuracy. Results revealed that all the innovative schemes can become attractive depending on water cost. The Water-Salt-Salt scheme produces the lowest levelized cost of sodium hydroxide at a concentration of 0.5 mol L-1, with a value of approximately 220 € ton-1 in favourable conditions. This suggests that in addition to environmental benefits, an economic improvement can be obtained selecting these innovative schemes.