Cost optimization of heat recovery steam generators and waste heat boilers : synthesis and design

The efficiency of combined cycles highly depends on the attention paid to the Heat Recovery Steam Generator (HRSG) during its overall design. Hence, engineers in charge of HRSG conception aim at optimizing heat transfer in order to recover the most possible heat from flue gas while limiting entropy production. To this end, an energy efficiency optimization is often conducted and generally provides, among many other thermodynamic parameters of the cycle studied, the optimal pressure levels for the HRSG, the mass flow of steam raised in each heat exchanger banks, etc. Such an optimization applied to Heat Recovery Steam Cycles has been introduced in February 2010 by Emanuele Martelli in his Ph.D Thesis entitled: “Numerical Optimization of Heat Recovery Steam Cycles for Highly Integrated Energy Systems” [1]. In this work, attention has mainly been given to thermodynamic efficiency considerations, since economic aspects were only taken into account by limiting the pinch point temperature differences and other key design parameters. Nowadays, cost considerations have nearly become as important as technical specifications for Engineering Procurement Construction companies (EPC), and are crucial for a project‘s successful execution. However, a sequential two-step approach is widely used in industry to come up with viable economic projects: indeed, the economic evaluation is often performed in the aftermath of the technical design and is therefore limited by the latter. In chemical industry, which often deals with complex chemical and energy facilities, engineers and researchers tried to include cost considerations in the technical design phase in order to get rid of the previously mentioned two-step strategy shortcoming. Hence, the well-known “Pinch Design Method” of Linnhoff has been broadened to deal with both area and energy targeting, and some new methodology based on Mixed Integer Non Linear Optimization Problems (MINLP) emerged with advent of even more powerful computers. This work proposes an adaptation of a MINLP model introduced by Yee and Grossmann in the 1990s for Heat Exchanger Networks, and summed up in two publications entitled: «Simultaneous optimization models for heat integration, I/II». Such a model provides a basis for a capital cost optimization of Waste Heat Recovery Units. With the purpose of proving the applicability of our methodology, the algorithm has been applied to three different units, including two HRSG of respectively 2 and 3 pressure levels, and one Waste Heat Boiler. Finally, part of this master thesis has been performed in the Foster Wheeler Energy Limited company, in Reading (UK), in the Fired Heater Department. This made it possible to compare our results with real industrial applications, and hence better understand to what extent our work fits EPC company requirements.