Study on Non Energy Saving Status Detection of Groundwater Heat Pump System Using Artificial Neural Network Method

摘要

Presents two types of characteristic data: basic characteristic parameters and index characteristic parameters for non energy saving status detection (NESSD) of groundwater heat pump (GWHP) system, establishes the relationship database between characteristic data and fault factors of NESSD. For three kinds of improving back propagation (BP) algorithms: Variable Learning Rate (VLR) BP algorithm, Scaled Conjugate Gradient (SCG) BP algorithm, and Levenberg-Marquardt (LM) BP algorithm, these various algorithms' comparative study had been conducted on the GWHP system's NESSD. The optimal algorithm among them is determined and the GWHP system's NESSD as cases studies can be carried out based on the most suitable BP algorithm.