Adaptive and predictive neural network models have been developed for a chiller and ice thermal storage tank of a central plant HVAC system. With relatively few input parameters, equipment performance is modeled without the use of first principles. Once trained, the models quickly predict output with relatively few calculations and self-calibrate, making them ideal for use with real systems that use adaptive and predictive controllers. The models are suitable for computer simulation or with actual systems.
Units: Dual
Citation: Symposium, ASHRAE Transactions, 1998, Vol 104, pt. 1A, San Francisco
Product Details
- Published:
- 1998
- Number of Pages:
- 8
- File Size:
- 1 file , 320 KB
- Product Code(s):
- D-7922