Optimization of Shell & Tube Heat Exchanger with Baffle-Cut, Industrial CFD Simulation

Description

Introduction

Before the advent of modern optimization software, engineers and scientists used classical optimization techniques to find optimal solutions. For better understanding, let`s assume an industrial and practical case study like a shell and tube heat exchanger, shown in the figures below. This shell and tube heat exchanger has 4 baffle cuts and 5 hot tubes in the center.

We aim to increase heat transfer while trying to minimize pressure drop. Notice that we can only change the mass flow of the inlet water, the baffle-cuts angle, and the diameter of hot tubes. Previously, in the classical optimization procedure, we simulated 10 designs, for example, and then chose the best-case scenario, but there was a problem.

There was more than 100 possible design, and we just picked the optimal design by testing just 10 of them! Thus, this couldn`t be called a real optimization! We`ve got an alternative now known as the Design of Experiment (DOE). Simply put, we can now choose the optimal design for all possible designs!!!

The basic idea of DOE is to vary the values of the design variables in a controlled manner and observe the resulting changes in the response variable. It involves several steps, which will be discussed In the following, by considering the same heat exchanger problem.

Project Description

A shell and tube heat exchanger is a type of heat exchanger that uses a shell (a large pressure vessel) to contain a bundle of tubes through which a fluid is circulated to exchange heat with another fluid. It is commonly used in industrial applications for heating or cooling large quantities of fluids.

In this project, we aim to find the optimal design of a shell and tube heat exchanger equipped with baffle cuts using ANSYS Workbench modules. The figure below shows a shell and tube heat exchanger with 4 baffle cuts and 5 hot tubes in the center.

The geometry is designed using ANSYS Design Modeler software. The computation domain was then divided into about a million tetrahedron cells via ANSYS Meshing software. The following figure shows the mesh grid:

Optimization

As mentioned, our target is the Optimization of Shell & Tube Heat Exchanger by increasing heat transfer while minimizing pressure drop. Numerical and equivalent output parameters should measure these objective functions. For example, the temperature of the fluid at the outlet can represent the heat transfer changes, and the pressure drop can be measured directly.

On the other hand, we can only change the mass flow of the inlet water, the baffle-cuts angle, and the diameter of the hot tubes. These three variables are known as input parameters/variables in optimizations. Also, we can change the input parameters in a specific range that are selected based on experience, doable, limitations, predictions, etc. For the current project, all of the data is summarized in the table below:

Input Variable	Range	Output
Baffle-cut angle (P1)	(-30°) – (+30°)	Outlet Temperature Pressure Drop
Tubes diameter (P2)	(10)-(25) cm
Inlet Mass Flow (P3)	(2)-(4) kg/s

The P1, P2 & P3 are marked in the figure below:

Optimization of Shell & Tube Heat Exchanger Methodology

All the optimization steps have been done by ANSYS Workbench software. This optimization has been done using Multi-Objective Genetic Aggregation. Also, the CCD algorithm is used as a Design of Experiments step, and Genetic Aggregation is used as an RSM step.

Results

The following table shows all the design points generated by the CCD method and the results extracted from executed simulations. By using these proposed 15 design points and its results, all other designs can be estimated. So we are looking for optimal design considering all possible designs, not just a limited number.

The second step is to generate response surfaces based on the Genetic aggregation method. The following figures are just some of the extracted response surfaces that can clearly show the effect of changing input parameters on objective functions.

For example, it can be seen that increasing the baffle-cut angle can increase the pressure drop in a specific range. The optimal point might be around point #1, marked in the figure in red color by just considering the pressure drop objective function. Furthermore, increasing the inlet’s mass flow could not unexpectedly enhance heat transfer.

So this time, point #2 may be a better design point. Many points like these can be candidates for optimal design depending on our prior criteria. Considering all the input parameters and the targets, there is a tradeoff that should be analyzed carefully in order to find the optimal design.

This process will be done in the third step of the optimization procedure using ANSYS Workbench software. After a quick analysis, the software gives us 3 best candidate points considering ALL OBJECTIVE FUNCTIONS and ALL INPUT VARIABLE RANGES. In other words, among all possible scenarios, not just limited ones.

As illustrated, using 17cm tubes, -30° baffle-cut angle, and 2 kg/s mass flow, the outlet temperature reaches 308.18K while experiencing just an 8.52 Pa pressure drop through the shell.