Status, Challenges, and Potential for Machine Learning in Understanding and Applying Heat Transfer Phenomena

2021 ◽  
Author(s):  
Matthew T. Hughes ◽  
Girish Kini ◽  
Srinivas Garimella
Keyword(s):  
Heat Transfer ◽  
Machine Learning ◽  
Large Scale ◽  
Computationally Efficient ◽  
Transfer Coefficients ◽  
Thermal Systems ◽  
Reduced Order ◽  
Heat Transfer Coefficients ◽  
Real Time Analysis ◽  
Transfer Problems

Abstract Machine learning (ML) offers a variety of techniques to understand many complex problems in different fields. The field of heat transfer, and thermal systems in general, are governed by complicated sets of governing physics that can be made tractable by reduced-order modeling, and by extracting simple trends from measured data. Therefore, ML algorithms can yield computationally efficient models for more accurate predictions or to generate robust optimization frameworks. This study reviews past and present efforts that use ML techniques in heat transfer from the fundamental level to full-scale applications, including the use of ML to build reduced-order models, predict heat transfer coefficients and pressure drop, real-time analysis of complex experimental data, and optimize large-scale thermal systems in a variety of applications. The appropriateness of different data-driven ML models in heat transfer problems is discussed. Finally, some of the imminent opportunities and challenges that the heat transfer community faces in this exciting and rapidly growing field are identified.

scholarly journals Heat Transfer in Rotating Serpentine Passages With Trips Normal to the Flow

1992 ◽  
Vol 114 (4) ◽  
pp. 847-857 ◽  
Author(s):  
J. H. Wagner ◽  
B. V. Johnson ◽  
R. A. Graziani ◽  
F. C. Yeh
Keyword(s):  
Heat Transfer ◽  
Turbine Blade ◽  
Large Scale ◽  
Flow Direction ◽  
Operating Conditions ◽  
Transfer Model ◽  
Transfer Coefficients ◽  
Flow Equations ◽  
Turbine Blade Cooling ◽  
Heat Transfer Coefficients

Experiments were conducted to determine the effects of buoyancy and Coriolis forces on heat transfer in turbine blade internal coolant passages. The experiments were conducted with a large-scale, multipass, heat transfer model with both radially inward and outward flow. Trip strips on the leading and trailing surfaces of the radial coolant passages were used to produce the rough walls. An analysis of the governing flow equations showed that four parameters influence the heat transfer in rotating passages: coolant-to-wall temperature ratio, Rossby number, Reynolds number, and radius-to-passage hydraulic diameter ratio. The first three of these four parameters were varied over ranges that are typical of advanced gas turbine engine operating conditions. Results were correlated and compared to previous results from stationary and rotating similar models with trip strips. The heat transfer coefficients on surfaces, where the heat transfer increased with rotation and buoyancy, varied by as much as a factor of four. Maximum values of the heat transfer coefficients with high rotation were only slightly above the highest levels obtained with the smooth wall model. The heat transfer coefficients on surfaces where the heat transfer decreased with rotation, varied by as much as a factor of three due to rotation and buoyancy. It was concluded that both Coriolis and buoyancy effects must be considered in turbine blade cooling designs with trip strips and that the effects of rotation were markedly different depending upon the flow direction.

scholarly journals Heat Transfer in Rotating Serpentine Passages With Trips Normal to the Flow

1991 ◽  
Author(s):  
J. H. Wagner ◽  
B. V. Johnson ◽  
R. A. Graziani ◽  
F. C. Yeh
Keyword(s):  
Heat Transfer ◽  
Turbine Blade ◽  
Large Scale ◽  
Flow Direction ◽  
Operating Conditions ◽  
Transfer Model ◽  
Transfer Coefficients ◽  
Flow Equations ◽  
Turbine Blade Cooling ◽  
Heat Transfer Coefficients

Experiments were conducted to determine the effects of buoyancy and Coriolis forces on heat transfer in turbine blade internal coolant passages. The experiments were conducted with a large scale, multi–pass, heat transfer model with both radially inward and outward flow. Trip strips on the leading and trailing surfaces of the radial coolant passages were used to produce the rough walls. An analysis of the governing flow equations showed that four parameters influence the heat transfer in rotating passages: coolant–to–wall temperature ratio, Rossby number, Reynolds number and radius–to–passage hydraulic diameter ratio. The first three of these four parameters were varied over ranges which are typical of advanced gas turbine engine operating conditions. Results were correlated and compared to previous results from stationary and rotating similar models with trip strips. The heat transfer coefficients on surfaces, where the heat transfer increased with rotation and buoyancy, varied by as much as a factor of four. Maximum values of the heat transfer coefficients with high rotation were only slightly above the highest levels obtained with the smooth wall model. The heat transfer coefficients on surfaces, where the heat transfer decreased with rotation, varied by as much as a factor of three due to rotation and buoyancy. It was concluded that both Coriolis and buoyancy effects must be considered in turbine blade cooling designs with trip strips and that the effects of rotation were markedly different depending upon the flow direction.

scholarly journals Heat Transfer in Rotating Serpentine Passages With Trips Skewed to the Flow

1994 ◽  
Vol 116 (1) ◽  
pp. 113-123 ◽  
Author(s):  
B. V. Johnson ◽  
J. H. Wagner ◽  
G. D. Steuber ◽  
F. C. Yeh
Keyword(s):  
Heat Transfer ◽  
Turbine Blade ◽  
Large Scale ◽  
Flow Direction ◽  
Operating Conditions ◽  
Transfer Model ◽  
Transfer Coefficients ◽  
Flow Equations ◽  
Turbine Blade Cooling ◽  
Heat Transfer Coefficients

Experiments were conducted to determine the effects of buoyancy and Coriolis forces on heat transfer in turbine blade internal coolant passages. The experiments were conducted with a large-scale, multipass, heat transfer model with both radially inward and outward flow. Trip strips, skewed at 45 deg to the flow direction, were machined on the leading and trailing surfaces of the radial coolant passages. An analysis of the governing flow equations showed that four parameters influence the heat transfer in rotating passages: coolant-to-wall temperature ratio, rotation number, Reynolds number, and radius-to-passage hydraulic diameter ratio. The first three of these four parameters were varied over ranges that are typical of advanced gas turbine engine operating conditions. Results were correlated and compared to previous results from similar stationary and rotating models with smooth walls and with trip strips normal to the flow direction. The heat transfer coefficients on surfaces, where the heat transfer decreased with rotation and buoyancy, decreased to as low as 40 percent of the value without rotation. However, the maximum values of the heat transfer coefficients with high rotation were only slightly above the highest levels previously obtained with the smooth wall model. It was concluded that (1) both Coriolis and buoyancy effects must be considered in turbine blade cooling designs with trip strips, (2) the effects of rotation are markedly different depending upon the flow direction, and (3) the heat transfer with skewed trip strips is less sensitive to buoyancy than the heat transfer in models with either smooth walls or normal trips. Therefore, skewed trip strips rather than normal trip strips are recommended and geometry-specific tests will be required for accurate design information.

Determination of Non-Uniform Heat Transfer Coefficients Around a Solid Aluminum Alloy Complex Shape During Quenching by Using an Inverse Method

2002 ◽  
Author(s):  
Pen-Chung Chen ◽  
Deborah A. Kaminski ◽  
Robert W. Messler
Keyword(s):  
Heat Transfer ◽  
Gas Turbine ◽  
Inverse Method ◽  
Flux Measurement ◽  
Transient Temperature ◽  
Transfer Coefficients ◽  
Heat Transfer Coefficients ◽  
Uniform Heat ◽  
The Cost ◽  
Transfer Problems

Gas turbine systems include complex heat transfer problems. Especially, the cooling efficiency is critical to the operation of gas turbine. In order to achieve the desired cooling condition, one needs to know the distribution of heat transfer on the components; however, the cost to implement a full-scale gas turbine test is tremendous. Therefore, many researchers used simplified models to acquire the test data; certain experiments can provide heat flux measurement, whereas other techniques can measure heat transfer coefficients. The direct measurement of heat transfer coefficients on the surface of components is extremely difficult. In such situations, the inverse method using transient temperature measurements taken within the part can be used to determine heat transfer coefficients. By combining experiments and numerical modeling, this presentation attempts to provide an effective and robust method to determine heat transfer coefficients on the part’s surface during cooling. Though the setting of the present paper is the quenching of a part, the technique presented is proposed for in-service heat load. To characterize the present situation, i.e., non-uniform heat transfer coefficients occurring during quenching, a unique methodology for employing inverse heat conduction was developed to obtain heat transfer coefficients from temperature responses. In conventional inverse approaches, the heat transfer coefficient is assumed to be uniform around the periphery, but this approach sometimes is unrealistic, especially for complex shaped parts. In this study, experimental data were used to find parameters in a heat transfer correlation, rather than to determine the coefficients directly. The resulting analysis provided an improved fit to measurements compared to conventional inverse approaches. The method developed was robust and is extendable to parts of arbitrary shape.

Use of Plate Thermometers for Better Estimate of Fire Development

2011 ◽  
Vol 82 ◽  
pp. 362-367 ◽  
Author(s):  
Alexandra Byström ◽  
Ulf Wickström ◽  
Milan Veljkovic
Keyword(s):  
Heat Transfer ◽  
Surface Temperature ◽  
Large Scale ◽  
Thermal Exposure ◽  
Alternative Methods ◽  
Gas Temperature ◽  
Transfer Coefficients ◽  
Heat Transfer Coefficients ◽  
Weighted Mean Temperature ◽  
Adiabatic Surface

The concept of Adiabatic Surface Temperature (AST) opens possibilities to calculate heat transfer to a solid surface based on one temperature instead of two as is needed when heat transfer by both radiation and convection must be considered. The Adiabatic Surface Temperature is defined as the temperature of a surface which cannot absorb or lose heat to the environment, i.e. a perfect insulator. Accordingly, the AST is a weighted mean temperature of the radiation temperature and the gas temperature depending on the heat transfer coefficients. A determining factor for introducing the concept of AST is that it can be measured with a cheap and robust method called the plate thermometer (PT), even under harsh fire conditions. Alternative methods for measuring thermal exposure under similar conditions involve water cooled heat flux meters that are in most realistic situations difficult to use and very costly and impractical. This paper presents examples concerning how the concept of AST can be used in practice both in reaction-to-fire tests and in large scale scenarios where structures are exposed to high and inhomogeneous temperature conditions.

Determination of heat transfer coefficients for large scale steel forgings quenched in polymer solutions

2018 ◽  
Vol 57 (1-3) ◽  
pp. 43
Author(s):  
Bradley P. Wynne ◽  
Rafael David Mercado Solis ◽  
Luis Adolfo Leduc Lezama ◽  
Edgar Ivan Saldana Garza ◽  
Jesus Mario Luna González
Keyword(s):  
Heat Transfer ◽  
Large Scale ◽  
Polymer Solutions ◽  
Transfer Coefficients ◽  
Heat Transfer Coefficients

The Effect of the Combustor-Turbine Slot and Midpassage Gap on Vane Endwall Heat Transfer

2011 ◽  
Vol 133 (4) ◽  
Author(s):  
Stephen P. Lynch ◽  
Karen A. Thole
Keyword(s):  
Heat Transfer ◽  
Large Scale ◽  
High Heat ◽  
Turbine Disk ◽  
Vortical Flow ◽  
Leakage Flow ◽  
Transfer Coefficients ◽  
Heat Transfer Coefficients ◽  
High Heat Transfer ◽  
The Individual

Turbine vanes are generally manufactured as single- or double-airfoil sections that are assembled into a full turbine disk. The gaps between the individual sections, as well as a gap between the turbine disk and the combustor upstream, provide leakage paths for relatively higher-pressure coolant flows. This leakage is intended to prevent ingestion of the hot combustion flow in the primary gas path. At the vane endwall, this leakage flow can interfere with the complex vortical flow present there and thus affect the heat transfer to that surface. To determine the effect of leakage flow through the gaps, heat transfer coefficients were measured along a first-stage vane endwall and inside the midpassage gap for a large-scale cascade with a simulated combustor-turbine interface slot and a midpassage gap. For increasing combustor-turbine leakage flows, endwall surface heat transfer coefficients showed a slight increase in heat transfer. The presence of the midpassage gap, however, resulted in high heat transfer near the passage throat where flow is ejected from that gap. Computational simulations indicated that a small vortex created at the gap flow ejection location contributed to the high heat transfer. The measured differences in heat transfer for the various midpassage gap flowrates tested did not appear to have a significant effect.

Heat Transfer in Rotating Serpentine Passages With Smooth Walls

1991 ◽  
Vol 113 (3) ◽  
pp. 321-330 ◽  
Author(s):  
J. H. Wagner ◽  
B. V. Johnson ◽  
F. C. Kopper
Keyword(s):  
Heat Transfer ◽  
Turbine Blade ◽  
Large Scale ◽  
Flow Direction ◽  
Vertical Tube ◽  
Operating Conditions ◽  
Transfer Model ◽  
Transfer Coefficients ◽  
Flow Equations ◽  
Heat Transfer Coefficients

Experiments were conducted to determine the effects of buoyancy and Coriolis forces on heat transfer in turbine blade internal coolant passages. The experiments were conducted with a large-scale, multipass, smooth-wall heat transfer model with both radially inward and outward flow. An analysis of the governing flow equations showed that four parameters influence the heat transfer in rotating passages: coolant-to-wall temperature ratio, Rossby number, Reynolds number, and radius-to-passage hydraulic diameter ratio. These four parameters were varied over ranges that are typical of advanced gas turbine engine operating conditions. It was found that both Coriolis and buoyancy effects must be considered in turbine blade cooling designs and that the effect of rotation on the heat transfer coefficients was markedly different depending on the flow direction. Local heat transfer coefficients were found to decrease by as much as 60 percent and increase by 250 percent from no-rotation levels. Comparisons with a pioneering stationary vertical tube buoyancy experiment showed reasonably good agreement. Correlation of the data is achieved employing dimensionless parameters derived from the governing flow equations.

Predicting System Performance and Estimating Parameters for Systems Burdened With Uncertainties and Noise Using Bayesian Inference

2008 ◽  
Author(s):  
A. F. Emery ◽  
D. Bardot
Keyword(s):  
Heat Transfer ◽  
Bayesian Inference ◽  
System Performance ◽  
Taylor Series Expansion ◽  
Substantial Effect ◽  
Simple Problem ◽  
Transfer Coefficients ◽  
Strong Function ◽  
Heat Transfer Coefficients ◽  
Transfer Problems

The precision of estimates of system performance and parameter estimation is often based upon the standard deviation obtained from the usual equation for the propagation of variances derived from a Taylor series expansion [1]. With increasing computing power, it is often suggested that the more complex Bayesian inference approach may yield improved estimates of the precision. The Bayesian approach has not been widely used in the heat transfer and fluid mechanics communities. The paper develops the necessary equations and applies them to two typical heat transfer problems. It is shown that, even for the simple problem of heat loss from a fin, that the predicted performance can be a strong function of relatively minor changes in the heat transfer coefficients or the thermal conductivity and as a consequence that the form of the parameter variability has a substantial effect.

Sign in / Sign up

Export Citation Format

Share Document