(661b) Thermo-Economic Optimization of Mechanical Vapor Recompression System for Shale Gas Produced Water Treatment
AIChE Annual Meeting
Thursday, November 1, 2018 - 12:55pm to 1:20pm
The present work presents a new optimization model for MVR process design with the objective of minimizing the total cost of produced water treatment. The optimization process utilizes a modified version of particle swarm optimization (PSO) algorithm which allows for relatively fast computation, of the design criteria of the MVR process while simultaneously minimizing the unit cost of produced water treatment. The primary model is developed through mass and energy balances for MVR system elements (including horizontal tube evaporator, centrifugal compressor, desuperheater, feed/distillate preheater, pumps, and mixers), and capital and operating cost equations. The operating condition constraint, energy balance, and heat transfer equations are added to the objective functions as proper penalty functions. Finally, the annual unit cost of treatment and penalty functions are optimized using PSO algorithm with iterative decision variables consisting of top brine temperature, temperature difference in the evaporator/condenser, temperature of the feed leaving the preheaters, and geometrical characteristics of the heat exchangers regarding their heat transfer coefficient calculations.
The developed optimization model is run for a hypothetical desalination plant with a capacity of 0.5 million gallons per day (MGD), which concentrates the produced water with 10% (100,000 mg/Liter) TDS to 30% TDS corresponding to saturation condition to make the exiting brine system capable of integration with downstream crystallizers. Sensitivity analysis is performed to identify the impact of input parameters such as brine salinity, feed flowrate, and feed salinity, on the energy consumption and unit cost of desalinating produced water. Feed salinity and brine flowrate are analyzed further to analyze the impact of these input variables on initial investment and equipment sizes. The results show that the proposed optimization model has to potential to work as a powerful tool in MVR design. Additionally, the final calculated costs show promising results for the application of MVR in the field of produced water treatment with unit costs of $2.2/m3(feed). The implications of these findings for sustainable management of produced water will be discussed.