.png){kind=icon}
返回
A new programmed method for retrofitting heat exchanger networks using graph machine learning
.png){kind=icon}
PRE
AI
DOI：10.1016/j.applthermaleng.2024.122427.png)
.png){kind=icon}
原文求助.png){kind=icon}
分享.png){kind=icon}
收藏
摘要
En 中文
Unsupervised graph machine learning provides a powerful framework for modeling and analyzing heat exchanger networks (HENs). This paper proposes a graph-based thermally guided path search (TGPS) method that systematically identifies and evaluates retrofit options to enhance the thermal performance of HENs. The method represents the HEN as a bipartite graph and uses automated algorithms to search for feasible heat integration paths. Thermodynamically inconsistent paths are filtered out based on temperature feasibility rules. The resulting retrofit options are evaluated using graphical metrics like betweenness centrality and cluster coefficients, as well as thermal performance indicators. A re-routing technique is introduced to address temperature mismatch issues for serial heat exchanger connections. When applied to a Kalina power cycle, the thermal efficiency of the optimum configuration is increased by 9.7%. This method is compared with both pinch analysis and the Energy Transfer Diagram approach, and it is thoroughly tested and verified for an ammonia-water absorption refrigeration cycle as well. This methodology provides an effective tool for retrofitting existing HENs while adhering to thermodynamic principles.
Keyword:
Graph theory
Graph machine learning
Heat exchanger network
Ammonia-water absorption refrigeration
Ammonia-water power cycle
期刊
IF:
6.9
论文数:
2.8W
被引数:
10.6W
机构
引用论文
分享收藏Analysis and optimization on a modified ammonia-water power cycle for more efficient power generation.png)
ENERGY
ENERGYIF9.4
PREAI
分享收藏Graphical analysis on internal heat recovery of a single stage ammonia-water absorption refrigeration systemENERGY
ENERGYIF9.4
PREAI
分享收藏分享收藏Estimation and sensitivity analysis of fouling resistance in phosphoric acid/steam heat exchanger using artificial neural networks and regression methodsSCIENTIFIC REPORTS
OAAI
SCIENTIFIC REPORTSIF3.9
OAAI分享收藏
查看更多内容