Development of ANN model for depth prediction of vertical ground heat exchanger. (February 2018)

• Record Type: Journal Article
• Authors: Chen, Shangyuan; Mao, Jinfeng; Chen, Fei; Hou, Pumin; Li, Yong
• Is Part Of: International journal of heat and mass transfer. Volume 117(2018)
• Page Start: 617
• Page End: 626
• Publication Date: 2018-02
• Subjects: Ground heat exchanger -- Numerical simulation -- Artificial neural network; Heat -- Transmission -- Periodicals; Mass transfer -- Periodicals; Chaleur -- Transmission -- Périodiques; Transfert de masse -- Périodiques; Electronic journals; 621.4022
• DOI: 10.1016/j.ijheatmasstransfer.2017.10.006
• Languages: English
• ISSNs: 0017-9310

Abstract

Highlights: A three-dimensional model of U-tube GHE is established and performed by CFD. Totally 140 cases are carried out in order to generate the sample set. Different neuron numbers and training algorithms are compared to obtain the optimal ANN. The Characteristic of GHE can be predicted by a three-layer network based on LM algorithm. Abstract: In the design of a ground heat exchanger (GHE), it is difficult to take all the factors into consideration. In this study, an artificial neural network (ANN) model has been developed, which can predict the depth of a vertical GHE according to the given design parameters. A three-dimensional model has been developed to obtain the training and testing data. Using the soil thermal conductivity, grout thermal conductivity, inlet flow, inlet water temperature, underground water velocity and heat flux as the input parameters, and the borehole depth as the output parameter, a three-layer ANN model has been developed. The performances of different training functions and neuron numbers have been investigated. The results show that the effects of the volumetric heat capacity and the porosity on the heat transfer of the GHE can be neglected, and the depth of a GHE can be predicted by the three-layer ANN model for given input parameters. The optimal ANN model uses the LM algorithm, and there are 10 neurons in the hidden layer.