Abstract:

The heat exchanger of crude oil petrochemical unit is often affected by fouling, which leads to the serious decline of heat exchanger performance and the decline of heat exchanger network performance. At the same time, due to the coupled relationship between heat exchangers, the performance decline of different heat exchangers has different effects on the change of heat exchanger network performance. In the past, cleaning decisions were mainly made based on the performance degradation of a single heat exchanger to a certain extent, resulting in the poor operation of the heat exchanger network. Therefore, a method of heat exchanger network cleaning decision-making based on intelligent prediction and mechanism was proposed in this paper. The intelligent prediction model was established for the operation data of heat exchangers to obtain the performance change trend of heat exchangers. Combined with the performance simulation model of heat exchangers network, the performance change trend of heat exchanger network was further obtained, so as to formulate the cleaning scheme from the perspective of the performance change of heat exchanger network. Based on the operation data of crude oil heat exchangers, the neural network prediction model was established, which has good prediction accuracy. Through the case study of heat exchange network of a crude oil distillation unit, when the performances of HE1, HE2 and HE5 heat exchangers decline at the same time, the annual additional utility energy consumption of heat exchanger network increases by 12.1%. Compared with the traditional cleaning scheme based on the performance of single heat exchanger, the annual additional cost of utility, loss cost and annual total cost of the cleaning scheme based on the performance degradation of heat exchanger network are reduced by 13.1%, 14.1% and 13.8%, respectively, while the cleaning number of times is only increased by 3.

Key words: heat exchanger, crude oil, fouling, cleaning, performance prediction, heat exchanger network

摘要：

原油石化装置的换热器常受到结垢影响，导致换热器性能衰退严重，换热网络性能也会随之衰退，同时换热器之间的耦合关系，导致不同换热器的性能衰退对换热网络整体性能变化的影响不同。以往的清洗决策主要是根据单台换热器性能衰退到一定程度来制定的，这会导致换热网络整体可能处在较差的运行状态。因此，本文提出一种基于智能预测和机理模型的换热网络清洗决策方法，基于换热器的运行数据建立智能预测模型，获得换热器性能变化趋势，结合换热网络的性能模拟模型，进一步获得换热网络的性能变化趋势，从而从换热网络整体性能变化的角度来制定清洗方案。研究表明，对收集到的原油换热器运行数据，建立神经网络预测模型，具有较好的预测精度。通过对原油精馏装置换热网络的案例分析，当HE1、HE2和HE5三台换热器同时发生性能衰退时，换热网络年度公用工程能耗费用将增加12.1%。与传统基于单台换热器性能衰退情况制定的清洗方案相比，从换热网络整体性能衰退角度制定的清洗方案，年度额外公用工程费用减少13.1%，损失费用减少14.1%，年度总费用减少13.8%，而清洗次数仅增加3台次。

关键词: 换热器, 原油, 结垢, 清洗, 性能预测, 换热网络