Discusses the application of an artificial neural network model to predict energy use in a complex institutional building without the need for a data acquisition system. A neural network has been applied to daily data collected manually by building personnel. A previously developed energy management system used linear regression and other statistical measures to develop formulas to predict the energy use for the building. The predictions and actual consumption are compared and discrepancies between predicted and observed energy use are investigated to determine whether they are due to unusual weather conditions, fluctuations in building use, malfunctioning systems, or other causes. Addresses the predictive performance issue. A comparison of the predictive ability of a neural network and a traditional statistical approach is presented. Neural network application issues are discussed along with results.
KEYWORDS: expert systems, calculating, energy consumption, buildings, comparing, performance, accuracy, energy conservation, energy management, controls.
Citation: ASHRAE Transactions 1993, Vol.99, pt.1
Product Details
- Published:
- 1993
- Number of Pages:
- 13
- File Size:
- 1 file , 1.1 MB
- Product Code(s):
- D-18122
