Performance simulation and economic optimization of the UCLA Cogeneration and Central Chiller Facility

A suite of models have been developed to simulate the functional performance of the UCLA Cogeneration and Central Chiller Facility with the primary objective being plant operating cost reduction. To achieve these ends, three principal categories of models have been developed. These categories include semi-empirical modeling of plant components and systems for performance simulation, load forecasting models for estimation of the campus demands for chilled water, steam and electricity and a probabilistic decision model to provide plant operators with informed, cost-optimal recommendations.; Performance modeling of the facility combines semi-empirical models of all the principal plant components and system level modeling of the primary plant systems. Campus demands for steam and chilled water represent fixed constraints that must be satisfied by facility production. These models are combined to estimate plant operating costs in response to environmental variables while maintaining campus demands.; Short term load forecasting models were developed to generate predictions of campus demand. These forecasting models predict the chilled water, steam, and electrical demands to reflect the characteristic daily and seasonal variations discovered in archived data. Fourier sine series approximations to the detrended monthly ensemble averages provide the short term forecasting needs.; A probabilistic decision model generates recommendations to plant operators for cost minimal discrete plant state decisions. Plant states are discrete combinations of chillers that may satisfy plant demands. The decision model generates recommendations for plant state configurations as a function of time. Future estimates of the plant operating costs used in the decision model, are the result of operating cost forecasts generated by the plant simulation models in response to the forecasted demands.; The modeling suite has been compiled into a prototype computer code using LabVIEW. The simulation code has a simple graphical user interface to provide an easy way for users to modify the varied input parameters, monitor solution progress and view the simulation results. Furthermore, the simulation platform allows for pursuit of continued analysis, developmental efforts and, ultimately, integration with plant transducers.