Water-Using Network Synthesis For Batch Process

Industrial fresh water usage and wastewater discharge have been a focus with the global water shortage and more and more badly polluted environment by various wastewater. Most of process industries consume huge amounts of freshwater and discharge almost the same huge amounts of wastewater. So how to optimize the structure of water-using network for process industries to reduce fresh water consumption and wastewater discharge is an urgent task. Process industries are generally classified into two types: continuous process and batch process. So far, water-using network synthesis (WNS) for continuous process has been studied thoroughly, but WNS for batch is very deficiently since the process characteristics for batch are time-dependent and dynamic.This work proposes a systematic framework for water-using network synthesis for batch process (BWNS). According to the number of contaminants contained in the water process, BWNS is classified into two types: water-using network synthesis for batch process with single-contaminant (S-BWNS) and water-using network synthesis for batch process with multiple contaminants (M-BWNS).Pinch analysis is a useful tool for heat exchanger network synthesis which is very similar to S-BWNS, since S-BWNS includes only one contaminant. So water pinch analysis is used in S-BWNS, which generally includes two steps: water target determination and water network design.During the process of water target determination, pure water surplus/deficit is defined, water cascade analysis (WCA) is presented to analyze all water-using units, which introduced water cascade table (WCT) to obtain water target, WCT is calculate easily by programming or electronic table such as EXCEL. Then a method of time-purity two dimensional diagram (TP-diagram) is developed to help design the water network. That is, time is taken as X-axis and purity is taken as Y-axis, and a rectangle is taken to be a water-using unit: the left and right border of which represent respectively the start time and the ending time of the unit, the up and down border represent respectively the inlet water purity and outlet water purity of the water-using unit. The letter in the rectangle stands for the number of the unit. In addition, most of the relations among water-using units are described in TP-diagram. Therefore the chance of water integration is easy to found, which makes the water network design quickly and efficiently.Batch water using process is classified into rigorous batch, half batch and mixed batch according to its time characteristic. In the paper the conceptions of storage and water cycle are clarified, and the definitions of local pinch and total pinch are presented. The author deduces that water target obtained by the total pinch analysis is no more than the summation of all local targets obtained by the local pinch analysis, total pinch is very instructional.During the process of water network design, two types of design rules for S-BWNS are presented: rules with storage and rules without storage. And also rules about the addition and combination of storages are presented. And the effect of the flowrate of water unit on the structure of water network is studied under a certain contaminant transfer load. Three cases are illustrated in the paper to show the efficiency the method. The results show the method is suitable to mass-transfer and no-mass-transfer process, half-batch, rigorous batch and mixed batch, and can reduce fresh water usage and wastewater discharge remarkably.In fact, there are always multiple contaminants in practical industries, so multi-contaminant water-using network for batch (M-BWNS) is more meaningful. M-BWNS is more complex than S-BWNS, so it is very troublesome and even impossible to solve by the method of diagram and table. A mathematic programming method is developed to solve M-BWNS in the paper, which eliminate time-characteristic by storages.A superstructure for M-BWNS is established firstly based on the setting-up rules proposed in the paper, and then it is described by a non linear programming (NLP) model, and the model is solved by a multi-step method. The main works include several parts as follows: (1) the definition of the key component for M-BWNS is clarified, and a deduction that the inner cycle is no-existed for M-BWNS is testified which means water reuse without regeneration for M-BWNS is meaningless; (2) the NLP model includes two layers: the inner layer and the exterior layer, the inner is used to determine the minimum freshwater of every water demand under the certain operation sequence, and the exterior is used to determine the minimum freshwater of the certain sequence; (3) the water source priority is presented, which reduces greatly the size of the NLP model; (4) the method to solve the NLP model is developed based on sequential concept, which includes two steps: firstly the optimum sequence and target network without time constraints is introduced; and then the number of needed cycles before the network steady is obtained by "the reverse number"; and also the stop criterion of calculation and the rules of storage clear-up and combination are presented; (5) the results of illustrated cases show the method developed in this work is a useful and efficient tool for multi-contaminant water-using synthesis for batch process, it can save a lot of freshwater, and can reduce wastewater discharge simultaneously.In the last part of the paper, M-BWNS with regeneration units is studied and a NLP model with one wastewater regeneration unit is established by a three-layer structure. The minimum fresh water, the minimum regenerated wastewater and the minimum regenerated contaminant mass loads are taken as the objectives representatively, a three segmental method is developed to solve the model. Two cases are solved by the method, the results show the method is valid to obtain the final steady water network, and the steps of calculation is easy, M-BWNS with regeneration units can save more fresh water and reduce more wastewater discharge.