(418a) Parameter Estimation for Flexible Fuel Energy Conversion Networks

We demonstrate the modeling of flexible energy conversion networks appropriate for use in campus scale and municipal scale utilities with complex energy requirements, fuel sources, and operational flexibility. We seek models that sufficiently capture the behavior of flexible fuel energy conversion networks work production with a minimal model. A previously reported class of bilinear models for estimating the efficiency of complex and flexible energy conversion networks offers a framework for such a model. Once an appropriate model is determined, this model may be used by the utility operator to determine financially optimal operating conditions and opportunities for financial and operational hedging.

In this paper, we use a previously introduced steady state bilinear model to model real power plants. This class of bilinear models incorporates first and second law principles from finite-time thermodynamics to predict energy conversion networks. We seek models that capture the behavior of work output over a range of heat inputs. These models need to contain enough complexity to accurately capture the work output behavior but more simplistic models require estimation of fewer parameters.

From a given energy conversion network model we compute a unique mapping from the decision variables and heat input to output work. We call this the input/output model of the energy conversion network. A given input/output model may have multiple network realizations. We demonstrated the realization of a one engine/three node and a one engine/four node model from plant data reported in a report published by the California Energy Commission.

Calibration of the model uses measured work output as a function of heat input. Typically this data are available in the form of heat rate curves. Given a particular model and data set, a data fitting procedure is developed to determine the parameters that give the best calibrated model for the data.

In summary, this paper presents three main results:

? Several options of potential energy conversion network models are reviewed and the resulting functional forms of the work output as rational functions of entropy flux with temperature parameters.

? A methodology is introduced for modeling flexible energy conversion networks using a class of bilinear models introduced in an earlier paper.

? Using heat rate curve data from real plants, the estimation of the parameters associated with the energy conversion network model is demonstrated.