Simultaneous synthesis of flexible heat exchanger networks

Abstract

In industry there is still lot of potential to make an energy system more efficient and thereby reduce the waste heat available. On the other hand there is an option to export the waste heat to another industry or to society. When the use of a heat exchanger network is considered for these tasks the optimization framework developed in this work can be implemented to calculate the cost of optimal investments.

This thesis presents a framework for generating flexible heat exchanger networks (HEN) over a specified range of variations in the flow rates and temperatures of the streams, so that the total annual costs (TAC) as a result of utility charges, exchanger areas and selection of matches are minimized. The proposed framework includes (i) an initialization stage to reduce the problem size, (ii) a multiperiod simultaneous MINLP model to synthesize a flexible HEN configuration, (iii) a multiperiod LP feasibility test model to check the operability and identify critical conditions which are to be included in the possible resolve stage of the MINLP model, and (iv) an NLP improvement model for further optimization by partly removing simplifications related to the MINLP model. This framework results in a HEN which can work in varying conditions without losing stream temperature targets and can keep an economically optimal energy integration.

This thesis also shows how the simplified superstructure presentation proposed by Yee and Grossmann (1990a) can be applied for generating flexible heat exchanger networks. Furthermore, this thesis presents a scheme which eliminates the modeling of bypasses, so that the nonlinear heat balances, binary variables, temperature variables and flow variables related to each bypass in the superstructure are no longer needed in the model. The elimination of bypass modeling, a stage-wise superstructure presentation and isothermal mixing assumption, make the MINLP model more robust and efficient to solve. Since this MINLP model is not solved until a problem is well prepared by the other parts of the developed optimization framework, the methodology presented in this thesis is applicable to solve industrial size grassroot design cases of flexible heat exchanger network problems.

Lastly, the proposed HEN synthesis strategy has been successively applied to two industrial problems where the industrial waste heat streams have been cooled down, forming a local and site level energy integration to gain savings in steam consumption and to avoid cooling tower investment. Both these problems represent the special case of correlated uncertain parameters, which here means that there is a relationship between uncertain parameters given in the stream data.