Performance assessment and modeling of an SWRO pilot plant with an energy recovery device under variable operating conditions

Abstract

Reverse osmosis (RO) is one of the most widespread desalination technologies in use today due to its good performance and reliability. Given that it is an energy intensive technology, using variable renewable energy sources (VRES) to power RO systems is an interesting option. Work with the RO system under variable operating conditions is one of the strategies that can be employed to take advantage of all the energy that is available at any given time from an off-grid renewable system. However, this will entail additional challenges in terms of, among other factors, plant maintenance and permeate production rate and quality. In grid-connected seawater RO (SWRO) desalination plants, energy recovery devices (ERD) are commonly used to increase energy efficiency performance. In these cases, the ERD usually operates under constant operating conditions. This work aims to assess the performance of an SWRO system with an ERD under widely variable operating conditions. The SWRO system has six membrane elements in pressure vessels. The ERD is a Pelton turbine connected to a generator to measure the energy produced by the turbine. An artificial neural network (ANN) based model was developed to estimate the performance of the SWRO-ERD system under variable operating conditions. According to the results, power savings of between 2.9 and 6.08 kW can be achieved for a wide range of operating conditions, allowing an increase in the produced permeate flux (Qp). The proposed ANN-based model is able to estimate Qp and permeate electrical conductivity with error intervals of 1.56 × 10−6 - 8.49 × 10−2 m3 h−1 and 8.33 × 10−5 - 31.06 μS cm−1, respectively. The experimental data and the developed model could help to obtain a better performance prediction of VRES-powered SWRO systems that are operating under variable operating conditions and with ERDs.