Experimental and machine learning-based predictions of thermohydraulic performance in gasketed plate heat exchangers with three corrugated plate geometries

Gasketed plate heat exchangers (GPHEs) are widely used in thermal systems, yet substantial discrepancies persist among published friction-factor correlations, even for nominally similar corrugated geometries. A key unresolved cause is plate pack breathing—elastic deformation driven by inlet-pressure asymmetry and tightening—which creates expanded and constricted channels with distinct hydraulic behavior. This work reports a systematic experimental investigation of three corrugation patterns (Chevron, Zigzag, and Four-Quadrant) tested at three tightening levels (0.98 A, 1.00 A, and 1.02 A) under controlled inlet-pressure differences. Local friction factors ( f ), overall heat-transfer conductance ( UA ), and Nusselt numbers ( Nu ) were determined independently for each channel to quantify breathing effects. Results show that friction factors are highly sensitive to deformation: depending on geometry and tightening level, differences on f- values between expanded and constricted channels reach up to 100%, explaining much of the scatter in existing f–Re correlations. In contrast, heat-transfer performance is comparatively robust, with tightening level-induced data dispersion in Nu limited to 3–10% (Chevron), 4–12% (Zigzag), and 3–16% (Four-Quadrant). Predictive models were developed using symbolic regression and a multilayer perceptron (MLP) neural network. The symbolic models achieved mean absolute percentage errors (MAPE) of 3–9%, corresponding to a reduction in prediction uncertainty of up to 50% compared with classical correlations that neglect tightening and inlet-pressure asymmetry. The MLP model showed superior performance, retaining MAPE ≈ 3.5% and strong generalization under Reynolds-range exclusion tests. This study clarifies the causes regarding friction-factor inconsistencies in GPHEs and provides validated predictive tools for operation under asymmetric or non-uniform tightening conditions. • Local friction factors measured in expanded and constricted channels of GPHEs. • Plate breathing shown as key cause of scatter in f–Re correlations. • Nusselt only mildly affected by tightening across all geometries. • Chevron plate shows best overall thermo-hydraulic performance. • Symbolic regression and ANN accurately predict GPHE behavior.

Authors

Institutions

Publication Details

Journal
Applied Thermal Engineering
Published
2026-03-04
DOI
https://doi.org/10.1016/j.applthermaleng.2026.130506
Primary Topic
Heat Transfer and Optimization
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
article

Experimental and machine learning-based predictions of thermohydraulic performance in gasketed plate heat exchangers with three corrugated plate geometries

K.V. Paiva, A.M.P. Gabardo, G.S.M. Martins, A.O. Stein et al.
Applied Thermal Engineering
Heat Transfer and Optimization
article

Experimental and machine learning-based predictions of thermohydraulic performance in gasketed plate heat exchangers with three corrugated plate geometries

K.V. Paiva, A.M.P. Gabardo, G.S.M. Martins, A.O. Stein, P.A. Jaskowiak, J.L.G. Oliveira
article en

Abstract

Gasketed plate heat exchangers (GPHEs) are widely used in thermal systems, yet substantial discrepancies persist among published friction-factor correlations, even for nominally similar corrugated geometries. A key unresolved cause is plate pack breathing—elastic deformation driven by inlet-pressure asymmetry and tightening—which creates expanded and constricted channels with distinct hydraulic behavior. This work reports a systematic experimental investigation of three corrugation patterns (Chevron, Zigzag, and Four-Quadrant) tested at three tightening levels (0.98 A, 1.00 A, and 1.02 A) under controlled inlet-pressure differences. Local friction factors ( f ), overall heat-transfer conductance ( UA ), and Nusselt numbers ( Nu ) were determined independently for each channel to quantify breathing effects. Results show that friction factors are highly sensitive to deformation: depending on geometry and tightening level, differences on f- values between expanded and constricted channels reach up to 100%, explaining much of the scatter in existing f–Re correlations. In contrast, heat-transfer performance is comparatively robust, with tightening level-induced data dispersion in Nu limited to 3–10% (Chevron), 4–12% (Zigzag), and 3–16% (Four-Quadrant). Predictive models were developed using symbolic regression and a multilayer perceptron (MLP) neural network. The symbolic models achieved mean absolute percentage errors (MAPE) of 3–9%, corresponding to a reduction in prediction uncertainty of up to 50% compared with classical correlations that neglect tightening and inlet-pressure asymmetry. The MLP model showed superior performance, retaining MAPE ≈ 3.5% and strong generalization under Reynolds-range exclusion tests. This study clarifies the causes regarding friction-factor inconsistencies in GPHEs and provides validated predictive tools for operation under asymmetric or non-uniform tightening conditions. • Local friction factors measured in expanded and constricted channels of GPHEs. • Plate breathing shown as key cause of scatter in f–Re correlations. • Nusselt only mildly affected by tightening across all geometries. • Chevron plate shows best overall thermo-hydraulic performance. • Symbolic regression and ANN accurately predict GPHE behavior.

Applied Thermal EngineeringVol. 293
Universidade Federal de Santa Catarina (BR)
Affordable and clean energy
Openalex Percentile: Top 36%
Heat Transfer and Optimization
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.