Intelligence modeling of the flow boiling heat transfer of nanorefrigerant for integrated energy system Provisionally Accepted

Songyuan Zhang^{1, 2} Yuexiwei Li^{1, 2} Zheng Xu^{1, 2} Lei Ma^{1, 2*} Yongjia Li^{1, 2*}

• ¹Faculty of Metallurgical and Mining, Kunming Metallurgy College, China
• ²Yunnan International Joint Research and Development Center of Green Energy Storage Material Industrial Chain coupled with Digital Twin Technology, Kunming Metallurgy College, Kunming 650033, China, China

To promote the application of nanorefrigerant in Organic Rankine Cycle (ORC) and Integrated Energy System (IES), a reliable model with simple structure and favorable accuracy for predicting the flow boiling heat transfer coefficient (HTC) of nanorefrigerant is essential. In this work, four intelligence models—the radial basis function (RBF), multilayer perceptron (MLP), least squares support vector machine (LSSVM), and adaptive neuro fuzzy inference system (ANFIS)—were developed to predict the flow boiling HTC using nanorefrigerants, based on 765 experimental samples. The performances of these artificial intelligence models were comprehensively evaluated through accuracy analysis, variation trend analysis, and sensitivity analysis. Results indicated that the comprehensive performance of the RBF model was superior than those of other intelligence models and the existing empirical models. The RBF model accurately captured the variation trend of the output as the input variables were varied. Meanwhile, the impact degrees of all input variables in decreasing order were nanoparticle concentration (φ), mass flux (G), thermal conductivity of nanoparticle (kp), and vapor quality (x).