Simultaneous Optimization Of Synthesis And Cleaning Schedule In Flexible Heat Exchanger Networks

Flexible heat exchanger network can overcome adverse effects resulting from operation conditions disturbances and external environment changes, and maintain regular service of heat recovery system. So on the basis of the investigation of heat exchanger network synthesis at fix opeation conditions, its flexible issue is paid a lot of attention widely by the relevant scholars in recent years. How to resist the negative effect of fouling resistance on heat transfer performance of heat exchanger also becomes one of the main parts of the investigation of flexible heat exchanger network. However, the most common method of taking account of fouling is adopted by means of fouling factor during the course of flexible heat exchanger network synthesis. Although flexible target of heat recovery system can be met, equipment investment is increased undoubtedly, and the economical efficiency deviates from the optimum value. In view of the above problem the paper provides a different way for simultaneous optimization of synthesis and cleaning schedule in flexible heat exchanger network. The research content includes the following five parts.(1) A novel method of flexible heat exchanger network synthesis is presented in this paper. The effect of the fouling process on heat transfer performance is taken into account at synthesis stage. Combined with the strategy of periodic cleaning exchanger, the flexibility constraints of heat exchanger networks can be be met. Thus the optimum economical efficiency is achieved by means of the simutaneous optimization of flexible heat exchanger network synthesis and cleaning schedule.(2) The optimization problems of both flexible heat exchanger network synthesis and cleaning schedule belong to combinational optimization, and they are classed as NP-hard problems. In order to reduce computational complexity, a two-stage method is proposed. First of all, the period of time during system service is divided into several stages depending on fouling process. And then flexible heat exchanger network synthesis based on pseudo-temperature enthalpy diagram method is performed at every stage. Further, the corresponding structure obtained at above every stage is incorporated into a twofold over-synthesis network, and the associated matches are represented as the optimum candidates. On this basis, the network topology and cleaning schedule are generated randomly, and further the dynamic simulation can be performed, so that the total utility consumption during period of system service can be computed by means of integral. Depending on the cleaning schedule, the periodic cleaning of exchanger is perfomed, thus its heat transfer performance is restored. From what has been done above, the economical function involving fixed investment, operational cost and cleaning cost can be calculated. With the aid of Genetic/Simulated annealing algorithm, the related variables are optimized, thus the simultaneous optimization of both synthesis and cleaning schedule can be implemented.(3) Discrete time representation was often used to optimize cleaning schedule of heat exchanger network in previous works. However, a novel method named continuous time representation is presented in this paper to divide the whole service period into unequal sub-time intervals. And depending on state transition matrix and device status matrix, the cleaning and operation actions can be alternately performed at boundary of every sub-time interval. The method of continuous time representation may reduce the number of optimization variables, thus the computational complexity can be decreased. The effectiveness of the proposed method can be illustrated by example 2.(4) In order to improve calculation efficiency at simultaneous optimization stage, the each member of match candidates is randomly chosen by the aid of predetermined acceptance probability to construct network. The probability value of every match depends on the frequency of existing in synthesized networks of preliminary synthesis stage. The method of predetermined acceptance probability can improve the initial values of the optimum solution.(5) For a better explanation of the proposed method in this paper, three numerical examples are provided. The total annualized cost of the proposed method in example 1 is decreased by 15% than that of only flexible heat exchanger network synthesis. The result illustrates that considering cleaning schedule at synthesis stage is more favorable. The second example is given to illustrate that the proposed method may be applied to solve large-scale problem of flexible heat exchanger network synthesis. And the last example is concerned with the analysis of energy consumption about demethanizer system in ethylene plant.