Indirect evaporative cooling maps of China: Optimal and quick performance identification based on a data-driven model

Wenchao Shi, Xiaochen Ma, Yu Gu, Yunran Min, Hongxing Yang

Research output: Journal article publicationJournal articleAcademic researchpeer-review

14 Citations (Scopus)


The data-driven models of various air conditioning (AC) systems have been developed because of the wider application of machine learning in the engineering field. Indirect evaporative cooler (IEC), known as one of the effective and environment-friendly AC devices, achieves the cooling purpose without using any types of mechanical compressors or chemical refrigerants. Recent studies on various IECs have been carried out in full swing with a large amount of valuable data produced. However, the data-driven model of the cross-flow IEC for sensible and total cooling is yet to be developed. In addition, by extracting the indoor cold exhaust air into the secondary air channel, the application range of an IEC can be extended, but so far the performance of IEC used in different regions has been rarely evaluated. In this study, an IEC model was established based on the artificial neural network (ANN), which was validated with on-site measurement results from a real engineering project. Combining the selected geometric size of IEC and various outdoor weather conditions into the IEC-ANN model, a case study was conducted to present the annual and seasonal IEC performance maps of China, and the optimal application regions could be determined. Results show that south China, east China, and middle China are more suitable to employ IEC for air treatment and energy saving. In south China, the greatest average temperature drop caused by the IEC is 4.52 ℃. The maximum cooling capacity can reach 5.74 kW, and it accounts for 30.1 % of the total cooling load. In the typical office building, the seasonal energy saving of the IEC with the given size is up to 3.64 kWh/m2, and the annual energy saving can reach 6.02 kWh/m2. In addition, the inference time of this IEC-ANN model was significantly shorter compared with a numerical model. Based on the quick prediction speed, the model can improve the working efficiency in the design stage of the engineering and may provide a swift response to guide the system operation.

Original languageEnglish
Article number116047
JournalEnergy Conversion and Management
Publication statusPublished - 15 Sept 2022


  • Air conditioning
  • Cooling map
  • Data-driven model
  • Indirect evaporative cooling
  • Optimal and quick identification

ASJC Scopus subject areas

  • Renewable Energy, Sustainability and the Environment
  • Nuclear Energy and Engineering
  • Fuel Technology
  • Energy Engineering and Power Technology


Dive into the research topics of 'Indirect evaporative cooling maps of China: Optimal and quick performance identification based on a data-driven model'. Together they form a unique fingerprint.

Cite this