Implementation of Adaptive Network-Based Fuzzy Inference for Hybrid Ground Source Heat Pump

This paper presents a new development of an Adaptive Network-based Fuzzy Inference System (ANFIS) for a Hybrid Ground Source Heat Pump (HGSHP). The HGSHP is equipped with a supplementary heat sink composter to process organic solid waste (OSW), utilizing excess hot air from the condensing unit to aerate the compost pile. The Fuzzy Logic Controller (FLC) was developed using data collected by effective sensors installed in the HGSHP system. The main objective is to control the water flow rate with a Variable Speed Drive (VSD) to improve overall system performance. The dataset for ANFIS has been created and trained using MATLAB® software, then implemented on a Raspberry Pi nano-computer with Python coding. This paper compares the performance of ANFIS with two different cases: ANFIS with Triangular Membership Function (TriMF) and ANFIS with Gaussian Membership Function (GaussMF). After implementing ANFIS with TriMF and GaussMF, the average COP during composter operation and system cooling significantly increased. In contrast, the HGSHP system power consumption is sufficiently reduced in both case studies. Moreover, ANFIS also benefits the composting process, as evidenced by the increase in composter operation time, and vice versa for system cooling time. Ultimately, the implementation of ANFIS can improve the HGSHP system performance in both the TriMF and GaussMF cases, with the TriMF case showing a significant improvement in the HGSHP system performance compared to the GaussMF case.