Energy-saving Diagnosis Of Ground Water-source Heat Pump System Based On Artificially Neural Network

At present, the building energy efficiency has become the national focus. Architectural applications of renewable energy sources have been encouraged. A number of architectural applications of renewable energy demonstration projects were built, and renewable energy construction model cities were established in order to increase the construction efforts. In the construction of renewable energy applications, ground water source heat pump system is widely used in China, due to high efficiency and low operation cost. However, because of lack of effective energy-saving diagnostic monitoring system, the actual operation often leads to system operation efficiency greatly reduced. This topic was put forward under such a background.Based on the existing energy-saving standards for civil underground water-source heat pump air-conditioning system, this paper determined indicators of energy status, according to "water-source heat pump " (GB/T19409-2003), " Design standard for energy efficiency of public building " (GB 50189-2005), and in conjunction with experts and the actual operation staff experience and the author's quantitative analysis. The relationship between the basic characteristics parameters, target characteristic parameters and non-energy factors was established. And then the digital generation rule of parameters was determined. According to the relationship between parameters and non-energy factors, this paper created the base model between the standard input of the characteristic parameters of the standard sample and system output of energy-saving for diagnosis. Based on artificial neural network technology the training learning, memory simulation, nonlinear approximation, etc., the groundwater source heat pump system network model of energy-saving diagnosis was established. The network was trained and compared by three kinds of neural network BP algorithms, namely, VLBP (variable learning rate back propagation) algorithm, SCG (scaled conjugate gradient) algorithm and the LM (levenberg-marquardt) algorithm, then a best algorithm was choose. Finally the basic model of the groundwater source heat pump system and energy-saving run-time diagnostic system was accomplished. The paper has provided an effective monitoring method for energy efficient operation.