Demineralization of water is the final treatment for production of high-quality boiler feed water in steam-dependent manufacturing as oil refineries. The process prevents metallurgical and efficiency problems in boilers by capturing dissolved ion in water until saturation or exhausting of equipment, when cleaning of ion exchange resins or reverse osmosis membranes regenerates these treatments for a new period of processing, the former within hours and the latter over months.
Used in high viscosity fluid movements, hydrogen production, and processing of raw material by promoting mass and heat transfer driving forces, steam also produces electric power in turbines in a direct way, as in powerhouse facilities, or indirectly in high to low pressure steam transformations. Due to inherent uncertainties in the demineralized water treatment facility related to ion concentration in captured freshwater, random flow distribution in parallel network, and amount and quality of condensate water from process field recycle, better scheduling predictions in the cleaning stage of saturated/exhausted equipment may prevent production bottlenecks and lower plant utilization by steam and electrical shortages as a consequence of boiler feed water unavailability.
Although optimization have been proposed for design of grassroot- and retrofit of water treatments as well as pinch analysis for recycle, reuse, and regeneration of water, to the best of our knowledge, there is no work for scheduling optimization of water systems excepted for those found in irrigation and water pumping problems. In these works, the uncertainty related to inorganic salt concentration variability is neglected by fixed contaminants concentration in the inlet-flow of water consumers, treatments, etc. Fixed contaminant recovery in water treatments or contamination ratio in water-using units are also commonplace; thus, reducing these optimization problems to linear (LP) or mixed-integer linear (MILP) models, creating a gap for their applicability to industrial processes. Recently, Yang et al. considers removal efficiency variation by using shortcut models for water treatment units to exploit trade-offs between their cost and removal efficiency, where is presented a modified Lagrangean-based decomposition algorithm in order to solve the resulting nonconvex mixed-integer nonlinear programming (MINLP) problem efficiently.
We propose a scheduling optimization for short-term regeneration of ion exchange resin beds, where these treatment units are modeled as continuous-process with 2-modes of operation (processing and regenerating) using hypothetical pools for controlling accumulated processing volume until the saturation of the resins, considering pool filling during processing and drawing during regenerating. The processing volume until resin saturation is updated every each regeneration cycle by using the past data gathered continuously in the field for each ion exchange resin vessel, where measurement of electrical conductivity in the outlet-flow of the treatment alerts the necessity of regeneration. The model is an MILP problem by considering continuously the processing volume until saturation updates to reflect the uncertainties on dissolved ion concentration in the water treatment feed and random flow distribution in the parallel network. In the case of a contaminant concentration modeled as a variable, the model becomes an MINLP problem, where complicated networks and operations can be very difficult to solve, even to a feasible point.
Super dimensioned facilities capable of feeding boiler feed water constantly to avoid its unavailability is possibly the responsible for the lack of studies in scheduling optimization of regeneration of ion exchange resins. An industrial example demonstrates the approach and the resulting process-of-work improvements in the hand-operated regeneration, where a timetable for opening/closing of valves to spare time of processing before saturation of resins as well as for their regeneration can guide operator’s teams in the field.
(1) Alva-Argez A, Vallianatos A, Kokossis A. A multi-contaminant transhipment model for mass exchange networks and wastewater minimisation problems. Comput Chem Eng, 1999, 23, 1439−1453.
(2) Galán B, Grossmann IE. Optimal design of distributed wastewater treatment networks. Ind Eng Chem Res, 1998, 37, 4036−4048.
(3) Karuppiah R, Grossmann IE. Global optimization for the synthesis of integrated water systems in chemical processes. Comput Chem Eng 2006, 30, 650−673.
(4) Linlin Y, Salcedo-Diaz R, Grossmann IE. Global optimization for the synthesis of integrated water systems in chemical processes. Ind Eng Chem Res, 2014, 30, 650−673.
(5) Faria DC, Bagajewicz MJ. Planning model for the design and/or retrofit of industrial water Systems. Ind Eng Chem Res, 2011, 50, 3788−3797.
(6) Khor CS, Shah N, Mahadzir S and Elkamel A. Optimisation of petroleum refinery water network systems retrofit incorporating reuse, regeneration and recycle strategies. Can J Chem Eng. 2012, 90, 137−143.
See more of this Group/Topical: Computing and Systems Technology Division