This project presents a detailed technical assessment of two vertical tube cooker (shell-and-tube heat exchanger) designs using Computational Fluid Dynamics (CFD). The analysis was conducted for Tanis, a manufacturer of candy production lines, with the objective of verifying and improving the thermal performance of the tube cooker, a critical process unit for heating viscous product fluids. The focus was on quantifying heat transfer efficiency, characterizing flow patterns—particularly leakage and mixing on the shell side—and recommending design improvements based on CFD results.
Problem Statement and Objectives
The tube cooker consists of a vertical shell containing a tube bundle through which steam flows, while the product fluid circulates on the shell side. Baffles are used to direct the product fluid across the tube bundle, maximizing heat transfer. However, unavoidable assembly tolerances between baffles and the shell introduce gaps that permit bypass (leakage) flows. These leakage currents can reduce the effective heat transfer by allowing portions of the fluid to avoid contact with the heated tubes.
The primary objectives of the CFD study were:
- To evaluate the flow pattern and mixing in the shell side as well as the actual heat transfer.
- To assess the impact of geometric modifications (baffle spacing, addition of inactive tubes) on these parameters.
- To provide actionable recommendations for design enhancement, grounded in simulation results.
CFD Model and Methodology
Geometry and Mesh
Both designs have the same cross-sectional area (152 mm x 310 mm) and a tube bundle of 196 active 10 mm tubes. The original cooker comprised 46 baffles (32 mm spacing), while the modified design had 18 baffle sections (80 mm spacing) and included additional inactive tubes to block specific bypass paths. A finite volume mesh with local refinements captured the complex flow and thermal boundary layers, totaling 19.8M cells for the original and 40.2M cells for the modified cooker.
Fluid Properties and Boundary Conditions
Simulations were performed for a viscous fluid at two temperature regimes (inlet at 60°C and 100°C, outlet at 130°C), with a mass flow rate of 2300 kg/h. The flow was assumed laminar due to low Reynolds numbers. No-slip and adiabatic boundary conditions were applied to the shell, baffles, and inactive tubes, while active tubes were isothermal at 130°C.
Assessment Metrics
Key metrics included:
- Flow path distribution (percentage of fluid in preferred vs. leakage streams)
- Sectional and total heat transfer
- Temperature distribution across shell-side sections
- Residence time of fluid particles
Results and Key Findings
Flow Distribution and Leakage
In the original design, a significant portion of the shell-side fluid bypassed active heat exchange surfaces. Only 23% of the fluid followed the preferred zigzag (B-stream) path, directly interacting with the tubes. Leakage flows—primarily through gaps at baffle cuts (E-stream) and along the shell (F-stream)—accounted for 60% of the total flow, with 32% specifically through the E-stream. The C-stream, a bypass at the shell edge, contributed an additional 18%.
The modified design, which increased baffle spacing and introduced inactive tubes to block the C-stream, shifted the flow distribution notably. The preferred B-stream rose to 40%, and the C-stream was reduced to 9%. Overall, the total zigzag flow (B + C) increased from 41% to 49%, while leakage flows dropped from 60% to 52%. This redistribution is attributed primarily to the increased baffle distance.
Heat Transfer Performance
Heat transfer performance was evaluated by considering the heat transfer coefficient per unit length. This 1D heat transfer coefficient was larger in the modified design, indicating enhanced overall performance compared to the original design. Mixing was poor in the original and modified design.
Impact of Activating Additional Tubes
A further simulation activated the ring of previously inactive 10 mm tubes in the modified cooker. This yielded a modest increase (approximately 2°C) in average cross-sectional fluid temperature, mainly in the first section where flow direction favors tube contact. While the effect was localized, it demonstrates the potential benefit of maximizing the number of active heat transfer surfaces, particularly in regions where flow is perpendicular to the tubes.
Residence Time Analysis
A further simulation activated the ring of previously inactive 10 mm tubes in the modified cooker. This yielded a modest increase (approximately 2°C) in average cross-sectional fluid temperature, mainly in the first section where flow direction favors tube contact. While the effect was localized, it demonstrates the potential benefit of maximizing the number of active heat transfer surfaces, particularly in regions where flow is perpendicular to the tubes.
Discussion
The CFD analysis demonstrated that shell-side flow distribution is critically sensitive to baffle spacing and the strategic placement of inactive tubes. Increasing baffle distance and blocking the C-stream successfully redirected more fluid through the active tube bundle, improving thermal performance. However, the persistence of leakage flows through fabrication tolerances remains a limiting factor, and poor mixing means that cold leakage streams are not sufficiently heated before exiting.
Activating additional tubes yields a minor local temperature increase where the liquid enters the cooker. In the remaining sections, the liquid flow is along the tubes instead of perpendicular, which is less effective to exchange heat from the tubes to the liquid.
Conclusions
The modified tube cooker design, featuring increased baffle spacing and C-stream blockage, outperforms the original in both flow distribution and heat transfer efficiency. Key findings include:
- Improved proportion of preferred shell-side flow (B-stream) from 23% to 40%.
- Reduction in leakage flows from 60% to 52%.
- Higher 1D film heat transfer coefficients, indicating more efficient thermal performance.
- Poor mixing in shell-side flows results in cold leakage streams, limiting overall effectiveness.
- Activating additional tubes provides a minor local increase in fluid temperature.
Recommendations
Based on the CFD results, the following recommendations are made for further design optimization:
- Reduce Tolerance Gaps: Although challenging from a fabrication standpoint, minimizing the 1 mm tolerance gap would directly reduce leakage flows.
- Opposed Nozzle Placement: Placing inlet and outlet nozzles on opposite sides could better utilize the shell-side flow and increase outlet temperature by avoiding direct short-circuiting of cold streams.
- Activate More Tubes: Maximizing the number of active tubes, especially in regions with perpendicular flow, will yield incremental improvements in heat transfer.
- Optimize Baffle Geometry: Further parametric studies on baffle spacing, cut, and shape are warranted to find the optimal balance between structural feasibility and thermal performance.
- Enhance Mixing: Modifications to baffle design or additional features to promote mixing could mitigate the impact of cold leakage streams.
- Block Unfavorable Leakage Streams: Consider adding inactive tubes or extending baffle cuts with a ring at the shell wall to block the C- and E-stream, based on CFD-guided design iterations.