HTRI’s continuing research in process heat transfer has produced changes to several methods. Summarized here are those that may affect the cases users run in Version 9.4.

 

Heat transfer methods for laminar single-phase tubes with twisted tape inserts

Tubes with internal twisted tapes promote heat transfer at the expense of pressure drop. For single-phase heat transfer predictions in laminar flow (Reynolds number ≤ 2000), previous versions of Xchanger Suite used the Hong-Bergles correlations (1976) for Prandtl numbers ≤ 200 and the Marner-Bergles correlations (1989) for Prandtl numbers ≥ 2000. For Prandtl numbers between 200 and 2000, a linear proration between these two correlations was used. Feedback from industry indicated that these correlations could lead to inaccurate heat transfer predictions in laminar flow.

Version 9.4 includes an improved Manglik-Bergles correlation (1993) for heat transfer in laminar flow. Users can expect a reduction in heat transfer coefficients at low Prandtl numbers (< 200) when at lower Reynolds numbers (< 1000) and an increase in coefficients at high Prandtl numbers (> 1000) that becomes more pronounced for higher Reynolds number (> 2000). The improved correlation considers natural convection and offers a smoother transition than the previous linear proration. The pressure drop methods for laminar flow are unchanged.

Additional reading: Design Manual (Sections B2.1.2.2 and B3.1.2.2)

 

Heat transfer methods for inline high-finned economizer bundles

Economizer applications often utilize deep bundles of high-finned tubes and induce turbulent flow regimes at normal operating conditions. Xchanger Suite utilizes methods from AC-14 for the prediction of heat transfer coefficients for all inline high-finned layouts. These methods apply a 40% multiplier to penalize the prediction of heat transfer coefficients in inline bundles compared to staggered bundles.

Review of the research and comparisons with other accepted methods identified that this penalty is too conservative for deep inline bundles of high-finned tubes in the turbulent flow regime. Heat transfer research data are needed to validate and improve the methods. An interim approach is applied to reduce the conservatism for bundles with 10 rows or more and Reynolds numbers exceeding 10000. Xace^® models of inline high-finned bundles with more than 10 rows should see a significant increase in the prediction of heat transfer coefficients at higher Reynolds numbers (> 10000).

Additional reading: AC-14, Design Manual (Sections B2.4 and B3.4)

 

Supercritical heat transfer methods

HTRI has an active Technical Operating Plan (TOP) project 610-2 to collect data near the critical pressure of mixtures and pure components and produce an applicable method. The heat transfer coefficient and mechanism change dramatically as pressure approaches the critical pressure, mainly due to significant changes in physical properties. Preliminary studies of open literature data for water and carbon dioxide indicate that HTRI single-phase methods overpredict heat transfer coefficients in the supercritical regime.

After collecting additional data, HTRI plans to develop a suitable method that will be applicable to all fluids, but in the interim, Version 9.4 includes a heat transfer suppression factor for tubeside coefficients where the fluid is either carbon dioxide or water in the supercritical regime. The program applies this method when appropriate composition and conditions are detected, or the user may enable the method on the Control / Methods panel.

The impact of this interim solution depends on how the fluid properties were generated and whether the user followed the advice given in TT-13. For cases which showed significant variation in calculated coefficients as the fluid moves through the critical point, users should expect the local coefficient to significantly decrease. Less impact is expected for cases where users followed HTRI’s guidance on modeling supercritical fluids given in TT-13.

As shown in the image below, a case study with water provides an example where the results are most sensitive. The overall tubeside heat transfer coefficient decreases approximately 30%. The process conditions in this example lie heavily in the affected region. Applications which have a larger operating range compared to this focused study should see less of an impact. HTRI continues to recommend the guidance in TT-13.

Additional reading: TT-13, S-ST-3-18, STG-22, and Design Manual (Section B3.1.1.7)

 

New modeling option for tube coatings

Tube coatings applied to the tube side or shell side of a plain tube provide barriers to corrosion or fouling. Coatings also impact the efficiency of a heat exchanger while providing benefits to maintenance and overall equipment life. The inclusion of a coating impacts the thermal and hydraulic performance due to an increase in overall thickness and the additional resistance of the material.

HTRI has published our first report from TOP project 1000-2 on heat exchanger coatings, and Version 9.4 includes an initial software implementation. Version 9.4 models both internal and external coatings through specification of the thickness of the coating layer and the thermal conductivity of the material. Currently, the software does not consider the weight or vibration potential of a coating, nor does it factor them into mechanical calculations, which are assumed to hold true for very thin layers of coating material (i.e., less than 0.3 mm total). The coating fields should not be used to model double-walled tubes.  The modeling option for double wall tubes will be implemented in a future version. 

Additional reading: Design Manual (Section B1.2.2.3), F-32-TR, guidebook (forthcoming)

 

Modeling option to accommodate common inlet piping in reboilers

Common piping refers to piping that is shared between multiple exchangers. Xchanger Suite interprets piping specified as the flow path from the source to a single exchanger. The software assumes the specified piping is duplicated whenever exchangers in parallel are specified. Many users were not aware of this assumption which led to incorrect inputs, resulting in inaccurate pressure drop for piping elements common to multiple exchangers.

Version 9.4 allows Xist^® to model common inlet piping through specification of one of three element types (Flow Split- Tee, Flow Split-Branch, and Flow Split-Manifold) and use of pressure drop methods from TPF-12. These elements are disabled for exchangers in parallel due the inability to model piping to each heat exchanger. Users must manually set the flow split for branch and manifold elements, and a future release will extend this functionality. Additional details on the use of these elements are provided in the software help.

Additional reading: TPF -12 and webinar on common piping (forthcoming)

 

Dow Heat Transfer Fluids added to HTRI Fluid Properties package

Select Dow Chemical Company fluids have been available for many years in a variety of property databases such as the VMGThermo^® database and the HTRI legacy database. However, users have noted some fluid property value inconsistencies or discrepancies from published data. See HTRI Known Issue 24041 regarding SYLTHERM™ in VMGThermo with APR.

The HTRI legacy database was based on research test data HTRI gathered decades ago and is available only in the Composition-based property option. Accuracy within this package may vary due to the limits of the data it is based on. The longstanding HTRI recommendation for addressing the Known Issue as well as in general for obtaining the best and most accurate fluid properties has always been to obtain them directly from Dow or another reputable source.

HTRI partnered with Dow Chemical Company to provide an option for users to generate properties using their own fluid correlations validated to match DOWTHERM™ and SYLTHERM technical datasheets. Available through the HTRI Fluid Properties (new) package to all HTRI members, this extension includes 11 components with proprietary correlations for property generation.

 

Calculation of cold-side momentum pressure drop/gain for tubeside vaporizers

Previously, the momentum pressure gain/loss from nozzle to header was calculated only for thermosiphon cases with piping specified due to the inherent sensitivity of these units. When piping was not specified, such as in a generic shell-and-tube heat exchanger, this gain/loss was not included. Version 9.4 calculates this momentum change for all tubeside vaporizers, regardless of piping specification.

Users will see the impact of this change in the first and last tubeside pressure drop increments for vaporization cases without piping specified, primarily when the ratio of the nozzle flow area to the total tube flow area is significantly different. This pressure change is more significant at the outlet header where more vapor is present. In cases with specified piping, no change is expected.

Additional reading: Design Manual (Section B6.4.6.3)

 

Improved continuous fin pressure drop method

Member feedback prompted investigation into the accuracy of the f-factor method for continuous fin options in Xist and Xace. CFD simulations for continuous fin bundles indicated that the methods produced overly conservative pressure drops for single-phase vapor flow. HTRI’s evaluation identified two open literature alternatives, discussed in Section B2.8 of the Design Manual, that improved predictions. More recently, HTRI took experimental measurements on a five-row tube bundle and our Scaled-down Air Cooler Unit (SACU) to validate the methods.

Version 9.4 includes an initial adaptation of the resulting friction factors that significantly reduces the error in predicted pressure drop for continuous fins. This implementation applies each of the two literature methods at different ranges of Reynolds number and prorates between the methods during transition. Future adjustments are expected due to continuing research, but the initial implementation shows a median error of less than 25% against HTRI’s test data. This approach resolves Known Issue 36527. Users should expect substantial improvement in outside pressure drop predictions of these bundles with single-phase vapor.

While the implementation in Version 9.4 improves pressure drop prediction for bundles that closely matched our research bundle, the initial implementation did have some shortcomings. See Known Issue 40692 for more information.

Additional reading: Design Manual (Section B2.8)