AWWA MTC64614

A survey was performed over four seawater reverse osmosis (RO) desalination plants to study the current status of boron rejection in the seawater desalination process. The survey found that boron rejection ranged from 65 to 85 %, with some plants producing water having boron concentration over 0.5 mg/L, a maximum contaminant level in drinking water suggested in the World Health Organization (WHO) Guidelines. While installation of boron reduction systems increase the costs of the process, several system design options are available to meet the same target water quality criteria with highly varying costs. The objective of this study was to develop an algorithm based on a deterministic process model and a stochastic cost model that can be used to select the system with the lowest overall costs and risk. A deterministic predictive model was first developed to simulate the performance of a spiral wound module in terms of boron rejection, salt (TDS) rejection and overall product recovery. The model was verified using a set of pilot-scale experiments performed under different conditions using synthetic seawaters that contained boron. This model was then modified to simulate the full-scale RO process performance under varying design and operating conditions to meet required water quality criteria in terms of boron rejection. It was further combined with a stochastic cost-estimation model that estimates net present value (NPV) of future costs, which was used as an overall performance measure. The input probability models for the stochastic cost model were developed by fitting the distribution functions to random data. The key factors with uncertainties that have a significant impact on cost analysis include membrane costs, electricity prices, inflation rates, interest rates and discount rates. Four design options were selected from ten available design options based on the expected cost (NPV) and risk (variance in cost) for process optimization based on the coupled model. Includes 11 references, tables, figures.