Research On Energy Consumption And Efficicncy Of Public Buildings In Shenzhen

It is necessary to analyse and evaluate the energy consumption and energy efficiency potentiality of public buildings in a special city or zone, before implementing energy efficiency retrofit projects for old public buildings or energy efficiency design for new public buildings. Based on the project of Shenzhen governmental building energy survey and retrofit, this paper discusses the methods of the public building energy survey and field testing and gives some suggestions, and analyses the testing method of chiller's energy efficiency ratio and its error. After several months'survey and field testing, about seventeen public buildings'energy consumption data and seven typical public buildings'energy distribution and equipment running general state data were collected. This paper introduces the main features and capabilities of a new building energy simulation software-EnergyPlus. Analyse the differences of design-day cooling load calculation by harmonic response simplified method and EnergyPlus simulation. The joint application of testing and simulation is analyzed and based on an actual office building test data, this paper compare these two methods. Then it analyses the comprehensive energy efficiency potentiality and presents a proper retrofit plan. Neural network, which simulates the frame, the character and the cognizing ability of nerve element cerebrum, handle many kinds of information. Feed-forward neural network has a wide applied prospect in the modeling of nonlinear system. Aiming at building energy use prediction, this paper chooses the energy-related input and output data of an energyplus example as the training data of a BP network model. The model is trained by Levenberg-Marquardt algorithm and has satisfying accuracy compared to the energyplus model. Then the paper designs a similar BP model based on the historical daily energy use records of an office building in Shenzhen.This model has three layers and considered main factors that influence building energy use. The inputs of model are the outdoor dry bulb temperature, wet bulb temperature ,diffuse solar, direct solar,time sequence and weighted schedule ,and the output is the daily energy use. With performing good forecasting accuracy, the method is generally suitable for long period building energy use forecasting and can meet the demands of engineering application. At the end of this paper, many energy efficiency measurements are discussed and compared with orthogonal test method to confirm their relative significance.