TY - JOUR
T1 - An artificial intelligence based-model for heat transfer modeling of 5G smart poles
AU - Khosravi, A.
AU - Laukkanen, T.
AU - Saari, K.
AU - Vuorinen, V.
N1 - Funding Information:
The authors would like to acknowledge Business Finland for their financial support for the LuxTurrim5G project. We also acknowledge Aalto University , Department of Mechanical Engineering , for the support of the project.
Publisher Copyright:
© 2021
PY - 2021/11/1
Y1 - 2021/11/1
N2 - The LuxTurrim5G project is built on integrating different types of sensors and equipment that have been integrated into light poles in order to build new data-driven services. One additional service could be to harvest the waste heat produced in the electrical devices in the pole. In this research, we developed an intelligent model for heat transfer modeling of 5G Smart Poles. The input parameters used to construct the model are latitude of the station (deg), ambient temperature (°C), inside airflow (m3/min) and time (h). These input parameters are employed to predict heat flow (W) and maximum plate temperature (°C) inside the utility box. The results show that the ANFIS-PSO model provides an accurate prediction of R-value >0.95 for the test data, which is close to the maximum theoretically value of 1. The results showed that for the small amount of latitude, the maximum heat flow and temperature of the inside air is not detected at noon and the radiation heat flow to the vertical cylinder is maximized between sunrise and noon as well as between noon and sunset. The model also demonstrated that for the northern conditions, the temperature levels of heat generated over 30 °C are limited.
AB - The LuxTurrim5G project is built on integrating different types of sensors and equipment that have been integrated into light poles in order to build new data-driven services. One additional service could be to harvest the waste heat produced in the electrical devices in the pole. In this research, we developed an intelligent model for heat transfer modeling of 5G Smart Poles. The input parameters used to construct the model are latitude of the station (deg), ambient temperature (°C), inside airflow (m3/min) and time (h). These input parameters are employed to predict heat flow (W) and maximum plate temperature (°C) inside the utility box. The results show that the ANFIS-PSO model provides an accurate prediction of R-value >0.95 for the test data, which is close to the maximum theoretically value of 1. The results showed that for the small amount of latitude, the maximum heat flow and temperature of the inside air is not detected at noon and the radiation heat flow to the vertical cylinder is maximized between sunrise and noon as well as between noon and sunset. The model also demonstrated that for the northern conditions, the temperature levels of heat generated over 30 °C are limited.
KW - 5G smart pole
KW - ANFIS
KW - Artificial intelligence
KW - Heat transfer modeling
KW - Waste heat
UR - http://www.scopus.com/inward/record.url?scp=85118494738&partnerID=8YFLogxK
U2 - 10.1016/j.csite.2021.101613
DO - 10.1016/j.csite.2021.101613
M3 - Article
AN - SCOPUS:85118494738
SN - 2214-157X
VL - 28
JO - Case Studies in Thermal Engineering
JF - Case Studies in Thermal Engineering
M1 - 101613
ER -