[论文解读] Cleaning Schedule Optimization of Heat Exchanger Networks Using Particle Swarm Optimization

Totok R. Biyanto, Sumitra Wira Suganda+4 more|arXiv (Cornell University)|Dec 2, 2015

Process Optimization and Integration参考文献 5被引用 5

一句话总结

本文提出一种粒子群优化（PSO）方法，用于确定炼油厂换热器网络（HENs）的最优清洗计划，以最小化能耗损失和维护成本。通过在44个月期间平衡热量回收、清洗费用和泵送成本，该方法在100次迭代后实现收敛，总节省达123.6万美元——占最大潜在节省的23%。

ABSTRACT

Oil refinery is one of industries that require huge energy consumption. The today technology advance requires energy saving. Heat integration is a method used to minimize the energy comsumption though the implementation of Heat Exchanger Network (HEN). CPT is one of types of Heat Exchanger Network (HEN) that functions to recover the heat in the flow of product or waste. HEN comprises a number of heat exchangers (HEs) that are serially connected. However, the presence of fouling in the heat exchanger has caused the decline of the performance of both heat exchangers and all heat exchanger networks. Fouling can not be avoided. However, it can be mitigated. In industry, periodic heat exchanger cleaning is the most effective and widely used mitigation technique. On the other side, a very frequent cleaning of heat exchanger can be much costly in maintenance and lost of production. In this way, an accurate optimization technique of cleaning schedule interval of heat exchanger is very essential. Commonly, this technique involves three elements: model to simulate the heat exchanger network, representative fouling model to describe the fouling behavior and suitable optimization algorithm to solve the problem of clening schedule interval for heat exchanger network. This paper describe the optimization of interval cleaning schedule of HEN within the 44-month period using PSO (particle swarm optimization). The number of iteration used to achieve the convergent is 100 iterations and the fitness value in PSO correlated with the amount of heat recovery, cleaning cost, and additional pumping cost. The saving after the optimization of cleaning schedule of HEN in this research achieved at $ 1.236 millions or 23% of maximum potential savings.

研究动机与目标

• 为解决换热器网络（HENs）中结垢问题，该问题会降低传热效率并增加运行成本。
• 通过优化HEN组件的清洗计划间隔，最小化总成本。
• 平衡竞争因素：热量回收、清洗成本以及结垢引起的额外泵送成本。
• 应用粒子群优化（PSO）对复杂真实工业HEN系统中的调度决策进行优化。
• 展示PSO在有限迭代次数内实现显著成本节约的有效性。

提出的方法

• 采用代表性结垢模型，模拟换热器随时间的渐进性退化。
• 应用PSO算法对44个月运行周期内的清洗计划进行优化。
• 将适应度函数定义为三项成本分量的加权和：热量回收损失、清洗成本和额外泵送成本。
• 优化过程运行100次迭代，以确保收敛至近似最优解。
• PSO算法演化出一组候选清洗计划，基于适应度函数评估每个方案。
• 最终解根据最低总成本选定，反映出能源效率提升和维护成本降低。

实验结果

研究问题

• RQ1 为使总运行成本最小化，换热器网络中换热器的最优清洗间隔是什么？
• RQ2 结垢动力学的整合如何影响HEN维护调度中的成本-效益权衡？
• RQ3 在真实工业HEN系统中，粒子群优化能在多大程度上降低能耗与维护成本？
• RQ4 热量回收、清洗成本与泵送成本如何相互作用以确定最优清洗计划？
• RQ5 对于此复杂调度问题，PSO是否能在有限次迭代内收敛至近似最优解？

主要发现

• 优化后的清洗计划在44个月期间实现总计123.6万美元的节省。
• 这相当于通过改进调度可实现的最大潜在节省的23%。
• PSO算法在100次迭代内完成收敛，表现出计算效率。
• 适应度函数有效平衡了热量回收、清洗成本与额外泵送成本。
• 所提方法通过确保及时清洗，显著减少了因结垢导致的能耗损失。
• 结果证实，PSO是解决HEN清洗计划问题的一种可行且高效的优化技术。