Optimal Design of Flexible Power Cycles Through Kriging-Based Surrogate Models

Abstract

The paper presents a novel technique to define the optimal design of a power cycle considering design and off-design performance. A Kriging-based surrogate model is developed in order to simulate the power cycle at design conditions, while decreasing the computational effort. For each design considered, the performance at relevant off-design points is additionally evaluated by means of specific off-design models, developed for the main components of the system. The combination of design and off-design models allows the optimization process to take into account the performance at a selected set of operating conditions. The resulting optimal design will, thus, be characterized by a high degree of flexibility, intended as the ability to work efficiently in the several modes of operations to which the plant will be subjected to. The presented technique was tested on a case study. The optimal design of an offshore combined cycle was evaluated by using a multi-objective approach, where the two objective functions to minimize were the cumulative CO2 emissions and the weight of the bottoming cycle. The resulting designs showed to outperform those defined by a standard optimization procedure, demonstrating the effectiveness of the novel design technique.