Radiation and High-Temperature Heating

The engineering problem

Thermal radiation transfers energy through electromagnetic waves. Its fourth-power temperature dependence makes it unusually important at furnace temperatures, while emissivity, geometry and participating gases limit simple blackbody calculations. Radiation is not inherently absent at low temperature; it is simply competing with other transfer paths.

Scope and learning outcomes

Work with absolute temperatures and emissivity, model gray-body exchanges, and examine thermal radiation in furnaces and ovens. Students should be able to select a governing equation, construct a measurable boundary condition, calculate a plausible output with its units, and explain a failure case before applying a correlation.

Concept map

5.1 Blackbody emission and emissivity

A real surface emits less than a blackbody when its hemispherical emissivity is below unity; a surface facing surroundings exchanges net radiation. A hot processing tray cools to an enclosed room. Compare a polished and oxidized surface using Kelvin, and explain why Celsius to the fourth power is invalid. Representative relationship: Qdot=epsilonsigmaA*(Ts^4−Tsur^4).

5.2 Two-surface gray radiation

A gray parallel-plate enclosure can be represented by radiative resistances that include both surface emissivities. Two large furnace panels exchange heat across a narrow gap. Investigate when the large-parallel-plate approximation becomes poor because of edge view factors. Representative relationship: Qdot=sigmaA(T1^4−T2^4)/(1/e1+1/e2−1).

5.3 Radiation shielding and furnaces

An intermediate low-emissivity shield partitions radiative resistance, but its benefit depends on thermal contact and geometry. A metal radiation shield is installed inside a furnace enclosure. State the ideal adiabatic-shield assumption and why gas convection can erode the calculated reduction. Representative relationship: Qdot≈sigmaA(T1^4−T2^4)/Rrad.

5.4 Radiative and convective interaction

Surface temperature is determined by simultaneous convection, radiation and conduction; fitting a single apparent h may conceal changes with temperature. A heated pipe is exposed to moving air in a production hall. Derive the sign of both contributions and examine ambient temperature and air-speed uncertainty. Representative relationship: Qdot=A[h(Ts−Tinf)+epsilonsigma(Ts^4−Tsur^4)].

A disciplined workflow for this unit

Start with a diagram of a process boundary, including fluid inlets, outlets, phase boundaries, hot and cold external surfaces and any energy storage. Record which values are measurements, which are assumed, and which are calculated. Write the conservation statement before inserting a transfer correlation. After obtaining a numerical result, test the limiting case: the heat flow should vanish with zero temperature difference in a passive linear resistance model, while doubling the wall area should increase conduction duty if all other conditions remain fixed. For nonlinear radiation a twofold temperature increase cannot be treated as a twofold radiation increase.

Report results with their uncertainty drivers. For a heated process tube, a thermocouple may measure external wall temperature but not internal wall temperature, and a flow meter may give volumetric rather than mass rate. An apparent discrepancy between measured and predicted duty may be the result of imperfect knowledge of heat capacities, uncontrolled bypass or fouling. If a correlation is used beyond the printed range, label the result an extrapolation instead of assigning spurious precision.

Unit investigation method

Use a blank process sheet with five sections: known properties, governing physics, calculation, validation, and operating decision. At minimum compare two physically plausible operating states and identify what extra measurement would most efficiently distinguish them. Do not rely on the number of digits printed by a simulation to argue that a real process is controlled to that precision.

Unit 5, Lesson 1: Blackbody emission and emissivity

Unit: Radiation and high-temperature heating. Core formula: Qdot=epsilonsigmaA*(Ts^4−Tsur^4). Suggested model: [Laboratory 10 — Gray surface thermal radiation](../labs/lab-10.html).

Why this problem appears in process plants

A hot processing tray cools to an enclosed room. A real surface emits less than a blackbody when its hemispherical emissivity is below unity; a surface facing surroundings exchanges net radiation. This topic is a decision about where energy travels, which mechanism controls its rate, and what evidence is necessary to justify a calculation. The same reported heater duty can correspond to distinct wall temperatures depending on geometry, flow and material properties. A credible engineer must therefore distinguish the energy rate demanded by a process from the thermal resistance that allows the equipment to deliver that rate.

A process operator experiences this distinction as an operating constraint rather than as an equation. When product throughput increases, an unchanged heating utility can be insufficient even when the external wall stays hot. A measurement campaign should record mass flow, inlet and outlet conditions, utility conditions, local wall temperatures when available, and the operating history that may affect fouling. Without these signals, increasing utility temperature may mask the actual cause of poor performance and worsen degradation.

Physical mechanism and assumptions

A real surface emits less than a blackbody when its hemispherical emissivity is below unity; a surface facing surroundings exchanges net radiation. The equation selected here, Qdot=epsilonsigmaA*(Ts^4−Tsur^4), summarizes a limited physical model. Every term must refer to a declared boundary, an appropriate unit system, and a physically relevant value of temperature or material property. Before using the equation, identify whether the system is steady or transient, whether the fluid undergoes phase change, and whether the heat-transfer coefficient is known from a defensible correlation or simply a fitted number.

The engineer's first drawing should show an inward or outward normal for the relevant surface and a clear sign convention. Heat transport is directed down a temperature driving force in a passive system, although a controlled refrigeration cycle can move heat from a colder source to a hotter sink by supplying work. A temperature sensor located in a mixing chamber is not necessarily the right measurement of the wall temperature immediately adjacent to a process fluid. A fixed temperature boundary also differs from an imposed heat-flux boundary and gives a different mathematical problem.

Derivation and dimensional reasoning

Start from the first-law energy balance, accumulation = net incoming enthalpy + supplied heat + work, with the appropriate sign conventions. Introduce the constitutive expression Qdot=epsilonsigmaA*(Ts^4−Tsur^4) only after identifying the relevant energy-transfer mechanism. If a rate equation gives watts, confirm that conductivity is in W/(m K), surface area in m² and temperature difference in kelvin for a planar slab. The analogous exercise for a convection coefficient uses W/(m² K). For radiation, temperatures must be absolute because emission depends on the fourth power of thermodynamic temperature.

Terms that look superficially similar can require fundamentally different models. A cylindrical pipe wall does not have a constant radial area; a nonlinear radiation exchange cannot be replaced by a temperature-independent resistance without specifying a linearization point. Correlations expressed with Reynolds and Prandtl numbers need fluid properties at the proper reference temperature and geometry. In a thermally developing flow, the local Nusselt number may differ from its fully developed average. Write these distinctions directly beside the calculation rather than after the final answer.

Worked investigation: gray surface thermal radiation

We use a numerical teaching model with the following documented inputs: surface temperature = 700; Tsur = 310; epsilon = 0.8; A = 3. The primary control surface temperature is measured in K. Other parameters are Tsur [K]; epsilon [fraction]; A [m²]. Using the model equation, the predicted net outward radiation is 31.4182 kW. This is an arithmetic result for assumed properties, not a measured industrial duty or a certified equipment rating.

Increase the primary control from 700 to 772 K while holding the other inputs fixed. The same model predicts 47.0816 kW. Compare the absolute difference (15.6634 kW) with the percentage difference, and interpret the sign using the physical mechanism. Does this change chiefly alter the driving force, the thermal resistance, the material storage, or the rate at which the process moves energy? Explain why interpreting the curve without checking the other controlled inputs could be misleading.

The control range is deliberately finite. At very large or small values a simplified correlation can violate its validity conditions even if the numerical program continues to return finite numbers. For this lab, the important limitation is: Small convex diffuse gray surface facing large isothermal enclosure, view factor≈1. Decide whether the default state is inside that limitation and state what kind of additional property data, geometry measurement, or experimental coefficient would be needed to increase confidence in the estimate.

Measurement strategy and uncertainty

Compare a polished and oxidized surface using Kelvin, and explain why Celsius to the fourth power is invalid. Select one variable from the worked model that could realistically be measured on a plant floor, one that would normally be inferred, and one that is liable to drift over months of operation. Suggest how each could be independently checked. Thermal couples need installation and dynamic-response consideration, while flow meters require correct fluid properties and calibration. A sensor can be accurate at its measurement point yet irrelevant to the heat-transfer boundary being modeled.

Develop a first-order sensitivity calculation. If the output Y depends on parameter p, estimate dY/dp from two nearby runs of the simulator. If the measured parameter has uncertainty u(p), its approximate output contribution is abs(dY/dp)*u(p). This is a local linear approximation; it does not characterize poorly known geometry, phase-change instability or systematic installation bias. Keep uncertainty in the output's physical units until presenting any final percentage. Where a temperature difference is small, uncertainty can become the dominant proportion of the inferred transfer rate.

Process decision and limitation

Compare a polished and oxidized surface using Kelvin, and explain why Celsius to the fourth power is invalid. An excellent model is still a screening tool if it ignores details that control real equipment performance. A preliminary calculation may support an experiment or identify the direction of a sensitivity, but it cannot substitute for a vendor curve, verified thermophysical properties, design margins, pressure-equipment requirements or an installed sensor audit. Present the conclusion as an evidence-backed statement with a named range of validity. Avoid claiming that a temperature or duty is safe solely because one steady-state expression has a positive numerical solution.

Independent worked case

[Read a fully contextualized numerical case: Kelvin conversion in radiation](../worked-cases/unit-05-case-01.html).

Exercises

1. Starting with Qdot=epsilonsigmaA*(Ts^4−Tsur^4), identify the independent variables and produce a dimensionally consistent derivation or explanation of each term. Give an example where using the wrong unit would change the answer by at least a factor of 100.
2. Repeat the worked experiment with the control variable reduced by 10%. Compute the new output, the percent change, and compare the result to a first-order sensitivity estimate. Explain whether a linear approximation is justified.
3. Compare a polished and oxidized surface using Kelvin, and explain why Celsius to the fourth power is invalid. Describe the minimum sensor set and a method for recording uncertainty in the field.
4. Construct an alternative physical model for the same plant scenario by changing one boundary condition. State which observation could distinguish the two models.
5. Design a sensitivity chart comparing two material-property assumptions. What input dominates the uncertainty at the baseline and why?

Reproducible laboratory

Run [Laboratory 10: Gray surface thermal radiation](../labs/lab-10.html), export 151 modeled points to CSV, and annotate where the correlation is appropriate. Cite the chosen property source and write an explicit statement about the limits of extrapolation. The laboratory provides a schematic response curve and numerical readout but cannot validate a design without measured plant data.

Unit 5, Lesson 2: Two-surface gray radiation

Unit: Radiation and high-temperature heating. Core formula: Qdot=sigmaA(T1^4−T2^4)/(1/e1+1/e2−1). Suggested model: [Laboratory 11 — Radiation between two plates](../labs/lab-11.html).

Why this problem appears in process plants

Two large furnace panels exchange heat across a narrow gap. A gray parallel-plate enclosure can be represented by radiative resistances that include both surface emissivities. This topic is a decision about where energy travels, which mechanism controls its rate, and what evidence is necessary to justify a calculation. The same reported heater duty can correspond to distinct wall temperatures depending on geometry, flow and material properties. A credible engineer must therefore distinguish the energy rate demanded by a process from the thermal resistance that allows the equipment to deliver that rate.

A process operator experiences this distinction as an operating constraint rather than as an equation. When product throughput increases, an unchanged heating utility can be insufficient even when the external wall stays hot. A measurement campaign should record mass flow, inlet and outlet conditions, utility conditions, local wall temperatures when available, and the operating history that may affect fouling. Without these signals, increasing utility temperature may mask the actual cause of poor performance and worsen degradation.

Physical mechanism and assumptions

A gray parallel-plate enclosure can be represented by radiative resistances that include both surface emissivities. The equation selected here, Qdot=sigmaA(T1^4−T2^4)/(1/e1+1/e2−1), summarizes a limited physical model. Every term must refer to a declared boundary, an appropriate unit system, and a physically relevant value of temperature or material property. Before using the equation, identify whether the system is steady or transient, whether the fluid undergoes phase change, and whether the heat-transfer coefficient is known from a defensible correlation or simply a fitted number.

The engineer's first drawing should show an inward or outward normal for the relevant surface and a clear sign convention. Heat transport is directed down a temperature driving force in a passive system, although a controlled refrigeration cycle can move heat from a colder source to a hotter sink by supplying work. A temperature sensor located in a mixing chamber is not necessarily the right measurement of the wall temperature immediately adjacent to a process fluid. A fixed temperature boundary also differs from an imposed heat-flux boundary and gives a different mathematical problem.

Derivation and dimensional reasoning

Start from the first-law energy balance, accumulation = net incoming enthalpy + supplied heat + work, with the appropriate sign conventions. Introduce the constitutive expression Qdot=sigmaA(T1^4−T2^4)/(1/e1+1/e2−1) only after identifying the relevant energy-transfer mechanism. If a rate equation gives watts, confirm that conductivity is in W/(m K), surface area in m² and temperature difference in kelvin for a planar slab. The analogous exercise for a convection coefficient uses W/(m² K). For radiation, temperatures must be absolute because emission depends on the fourth power of thermodynamic temperature.

Terms that look superficially similar can require fundamentally different models. A cylindrical pipe wall does not have a constant radial area; a nonlinear radiation exchange cannot be replaced by a temperature-independent resistance without specifying a linearization point. Correlations expressed with Reynolds and Prandtl numbers need fluid properties at the proper reference temperature and geometry. In a thermally developing flow, the local Nusselt number may differ from its fully developed average. Write these distinctions directly beside the calculation rather than after the final answer.

Worked investigation: radiation between two plates

We use a numerical teaching model with the following documented inputs: hot temperature = 800; Tc = 350; e1 = 0.75; e2 = 0.65; A = 2.5. The primary control hot temperature is measured in K. Other parameters are Tc [K]; e1 [fraction]; e2 [fraction]; A [m²]. Using the model equation, the predicted net plate radiation is 29.8843 kW. This is an arithmetic result for assumed properties, not a measured industrial duty or a certified equipment rating.

Increase the primary control from 800 to 880 K while holding the other inputs fixed. The same model predicts 44.2811 kW. Compare the absolute difference (14.3968 kW) with the percentage difference, and interpret the sign using the physical mechanism. Does this change chiefly alter the driving force, the thermal resistance, the material storage, or the rate at which the process moves energy? Explain why interpreting the curve without checking the other controlled inputs could be misleading.

The control range is deliberately finite. At very large or small values a simplified correlation can violate its validity conditions even if the numerical program continues to return finite numbers. For this lab, the important limitation is: Large diffuse gray parallel plates, view factor≈1, no participating gas. Decide whether the default state is inside that limitation and state what kind of additional property data, geometry measurement, or experimental coefficient would be needed to increase confidence in the estimate.

Measurement strategy and uncertainty

Investigate when the large-parallel-plate approximation becomes poor because of edge view factors. Select one variable from the worked model that could realistically be measured on a plant floor, one that would normally be inferred, and one that is liable to drift over months of operation. Suggest how each could be independently checked. Thermal couples need installation and dynamic-response consideration, while flow meters require correct fluid properties and calibration. A sensor can be accurate at its measurement point yet irrelevant to the heat-transfer boundary being modeled.

Develop a first-order sensitivity calculation. If the output Y depends on parameter p, estimate dY/dp from two nearby runs of the simulator. If the measured parameter has uncertainty u(p), its approximate output contribution is abs(dY/dp)*u(p). This is a local linear approximation; it does not characterize poorly known geometry, phase-change instability or systematic installation bias. Keep uncertainty in the output's physical units until presenting any final percentage. Where a temperature difference is small, uncertainty can become the dominant proportion of the inferred transfer rate.

Process decision and limitation

Investigate when the large-parallel-plate approximation becomes poor because of edge view factors. An excellent model is still a screening tool if it ignores details that control real equipment performance. A preliminary calculation may support an experiment or identify the direction of a sensitivity, but it cannot substitute for a vendor curve, verified thermophysical properties, design margins, pressure-equipment requirements or an installed sensor audit. Present the conclusion as an evidence-backed statement with a named range of validity. Avoid claiming that a temperature or duty is safe solely because one steady-state expression has a positive numerical solution.

Independent worked case

[Read a fully contextualized numerical case: Parallel plate radiation denominator](../worked-cases/unit-05-case-02.html).

Exercises

1. Starting with Qdot=sigmaA(T1^4−T2^4)/(1/e1+1/e2−1), identify the independent variables and produce a dimensionally consistent derivation or explanation of each term. Give an example where using the wrong unit would change the answer by at least a factor of 100.
2. Repeat the worked experiment with the control variable reduced by 10%. Compute the new output, the percent change, and compare the result to a first-order sensitivity estimate. Explain whether a linear approximation is justified.
3. Investigate when the large-parallel-plate approximation becomes poor because of edge view factors. Describe the minimum sensor set and a method for recording uncertainty in the field.
4. Construct an alternative physical model for the same plant scenario by changing one boundary condition. State which observation could distinguish the two models.
5. Design a sensitivity chart comparing two material-property assumptions. What input dominates the uncertainty at the baseline and why?

Reproducible laboratory

Run [Laboratory 11: Radiation between two plates](../labs/lab-11.html), export 151 modeled points to CSV, and annotate where the correlation is appropriate. Cite the chosen property source and write an explicit statement about the limits of extrapolation. The laboratory provides a schematic response curve and numerical readout but cannot validate a design without measured plant data.

Unit 5, Lesson 3: Radiation shielding and furnaces

Unit: Radiation and high-temperature heating. Core formula: Qdot≈sigmaA(T1^4−T2^4)/Rrad. Suggested model: [Laboratory 12 — Ideal radiation shield](../labs/lab-12.html).

Why this problem appears in process plants

A metal radiation shield is installed inside a furnace enclosure. An intermediate low-emissivity shield partitions radiative resistance, but its benefit depends on thermal contact and geometry. This topic is a decision about where energy travels, which mechanism controls its rate, and what evidence is necessary to justify a calculation. The same reported heater duty can correspond to distinct wall temperatures depending on geometry, flow and material properties. A credible engineer must therefore distinguish the energy rate demanded by a process from the thermal resistance that allows the equipment to deliver that rate.

A process operator experiences this distinction as an operating constraint rather than as an equation. When product throughput increases, an unchanged heating utility can be insufficient even when the external wall stays hot. A measurement campaign should record mass flow, inlet and outlet conditions, utility conditions, local wall temperatures when available, and the operating history that may affect fouling. Without these signals, increasing utility temperature may mask the actual cause of poor performance and worsen degradation.

Physical mechanism and assumptions

An intermediate low-emissivity shield partitions radiative resistance, but its benefit depends on thermal contact and geometry. The equation selected here, Qdot≈sigmaA(T1^4−T2^4)/Rrad, summarizes a limited physical model. Every term must refer to a declared boundary, an appropriate unit system, and a physically relevant value of temperature or material property. Before using the equation, identify whether the system is steady or transient, whether the fluid undergoes phase change, and whether the heat-transfer coefficient is known from a defensible correlation or simply a fitted number.

The engineer's first drawing should show an inward or outward normal for the relevant surface and a clear sign convention. Heat transport is directed down a temperature driving force in a passive system, although a controlled refrigeration cycle can move heat from a colder source to a hotter sink by supplying work. A temperature sensor located in a mixing chamber is not necessarily the right measurement of the wall temperature immediately adjacent to a process fluid. A fixed temperature boundary also differs from an imposed heat-flux boundary and gives a different mathematical problem.

Derivation and dimensional reasoning

Start from the first-law energy balance, accumulation = net incoming enthalpy + supplied heat + work, with the appropriate sign conventions. Introduce the constitutive expression Qdot≈sigmaA(T1^4−T2^4)/Rrad only after identifying the relevant energy-transfer mechanism. If a rate equation gives watts, confirm that conductivity is in W/(m K), surface area in m² and temperature difference in kelvin for a planar slab. The analogous exercise for a convection coefficient uses W/(m² K). For radiation, temperatures must be absolute because emission depends on the fourth power of thermodynamic temperature.

Terms that look superficially similar can require fundamentally different models. A cylindrical pipe wall does not have a constant radial area; a nonlinear radiation exchange cannot be replaced by a temperature-independent resistance without specifying a linearization point. Correlations expressed with Reynolds and Prandtl numbers need fluid properties at the proper reference temperature and geometry. In a thermally developing flow, the local Nusselt number may differ from its fully developed average. Write these distinctions directly beside the calculation rather than after the final answer.

Worked investigation: ideal radiation shield

We use a numerical teaching model with the following documented inputs: shield emissivity = 0.15; Th = 900; Tc = 350; e1 = 0.85; e2 = 0.8; A = 2. The primary control shield emissivity is measured in fraction. Other parameters are Th [K]; Tc [K]; e1 [fraction]; e2 [fraction]; A [m²]. Using the model equation, the predicted shielded radiant duty is 5.28386 kW. This is an arithmetic result for assumed properties, not a measured industrial duty or a certified equipment rating.

Increase the primary control from 0.15 to 0.2828 fraction while holding the other inputs fixed. The same model predicts 9.69578 kW. Compare the absolute difference (4.41192 kW) with the percentage difference, and interpret the sign using the physical mechanism. Does this change chiefly alter the driving force, the thermal resistance, the material storage, or the rate at which the process moves energy? Explain why interpreting the curve without checking the other controlled inputs could be misleading.

The control range is deliberately finite. At very large or small values a simplified correlation can violate its validity conditions even if the numerical program continues to return finite numbers. For this lab, the important limitation is: One thin ideal adiabatic shield, equal areas, no gas radiation or convection. Decide whether the default state is inside that limitation and state what kind of additional property data, geometry measurement, or experimental coefficient would be needed to increase confidence in the estimate.

Measurement strategy and uncertainty

State the ideal adiabatic-shield assumption and why gas convection can erode the calculated reduction. Select one variable from the worked model that could realistically be measured on a plant floor, one that would normally be inferred, and one that is liable to drift over months of operation. Suggest how each could be independently checked. Thermal couples need installation and dynamic-response consideration, while flow meters require correct fluid properties and calibration. A sensor can be accurate at its measurement point yet irrelevant to the heat-transfer boundary being modeled.

Develop a first-order sensitivity calculation. If the output Y depends on parameter p, estimate dY/dp from two nearby runs of the simulator. If the measured parameter has uncertainty u(p), its approximate output contribution is abs(dY/dp)*u(p). This is a local linear approximation; it does not characterize poorly known geometry, phase-change instability or systematic installation bias. Keep uncertainty in the output's physical units until presenting any final percentage. Where a temperature difference is small, uncertainty can become the dominant proportion of the inferred transfer rate.

Process decision and limitation

State the ideal adiabatic-shield assumption and why gas convection can erode the calculated reduction. An excellent model is still a screening tool if it ignores details that control real equipment performance. A preliminary calculation may support an experiment or identify the direction of a sensitivity, but it cannot substitute for a vendor curve, verified thermophysical properties, design margins, pressure-equipment requirements or an installed sensor audit. Present the conclusion as an evidence-backed statement with a named range of validity. Avoid claiming that a temperature or duty is safe solely because one steady-state expression has a positive numerical solution.

Independent worked case

[Read a fully contextualized numerical case: Shield effectiveness](../worked-cases/unit-05-case-03.html).

Exercises

1. Starting with Qdot≈sigmaA(T1^4−T2^4)/Rrad, identify the independent variables and produce a dimensionally consistent derivation or explanation of each term. Give an example where using the wrong unit would change the answer by at least a factor of 100.
2. Repeat the worked experiment with the control variable reduced by 10%. Compute the new output, the percent change, and compare the result to a first-order sensitivity estimate. Explain whether a linear approximation is justified.
3. State the ideal adiabatic-shield assumption and why gas convection can erode the calculated reduction. Describe the minimum sensor set and a method for recording uncertainty in the field.
4. Construct an alternative physical model for the same plant scenario by changing one boundary condition. State which observation could distinguish the two models.
5. Design a sensitivity chart comparing two material-property assumptions. What input dominates the uncertainty at the baseline and why?

Reproducible laboratory

Run [Laboratory 12: Ideal radiation shield](../labs/lab-12.html), export 151 modeled points to CSV, and annotate where the correlation is appropriate. Cite the chosen property source and write an explicit statement about the limits of extrapolation. The laboratory provides a schematic response curve and numerical readout but cannot validate a design without measured plant data.

Unit 5, Lesson 4: Radiative and convective interaction

Unit: Radiation and high-temperature heating. Core formula: Qdot=A[h(Ts−Tinf)+epsilonsigma(Ts^4−Tsur^4)]. Suggested model: [Laboratory 10 — Gray surface thermal radiation](../labs/lab-10.html).

Why this problem appears in process plants

A heated pipe is exposed to moving air in a production hall. Surface temperature is determined by simultaneous convection, radiation and conduction; fitting a single apparent h may conceal changes with temperature. This topic is a decision about where energy travels, which mechanism controls its rate, and what evidence is necessary to justify a calculation. The same reported heater duty can correspond to distinct wall temperatures depending on geometry, flow and material properties. A credible engineer must therefore distinguish the energy rate demanded by a process from the thermal resistance that allows the equipment to deliver that rate.

A process operator experiences this distinction as an operating constraint rather than as an equation. When product throughput increases, an unchanged heating utility can be insufficient even when the external wall stays hot. A measurement campaign should record mass flow, inlet and outlet conditions, utility conditions, local wall temperatures when available, and the operating history that may affect fouling. Without these signals, increasing utility temperature may mask the actual cause of poor performance and worsen degradation.

Physical mechanism and assumptions

Surface temperature is determined by simultaneous convection, radiation and conduction; fitting a single apparent h may conceal changes with temperature. The equation selected here, Qdot=A[h(Ts−Tinf)+epsilonsigma(Ts^4−Tsur^4)], summarizes a limited physical model. Every term must refer to a declared boundary, an appropriate unit system, and a physically relevant value of temperature or material property. Before using the equation, identify whether the system is steady or transient, whether the fluid undergoes phase change, and whether the heat-transfer coefficient is known from a defensible correlation or simply a fitted number.

The engineer's first drawing should show an inward or outward normal for the relevant surface and a clear sign convention. Heat transport is directed down a temperature driving force in a passive system, although a controlled refrigeration cycle can move heat from a colder source to a hotter sink by supplying work. A temperature sensor located in a mixing chamber is not necessarily the right measurement of the wall temperature immediately adjacent to a process fluid. A fixed temperature boundary also differs from an imposed heat-flux boundary and gives a different mathematical problem.

Derivation and dimensional reasoning

Start from the first-law energy balance, accumulation = net incoming enthalpy + supplied heat + work, with the appropriate sign conventions. Introduce the constitutive expression Qdot=A[h(Ts−Tinf)+epsilonsigma(Ts^4−Tsur^4)] only after identifying the relevant energy-transfer mechanism. If a rate equation gives watts, confirm that conductivity is in W/(m K), surface area in m² and temperature difference in kelvin for a planar slab. The analogous exercise for a convection coefficient uses W/(m² K). For radiation, temperatures must be absolute because emission depends on the fourth power of thermodynamic temperature.

Terms that look superficially similar can require fundamentally different models. A cylindrical pipe wall does not have a constant radial area; a nonlinear radiation exchange cannot be replaced by a temperature-independent resistance without specifying a linearization point. Correlations expressed with Reynolds and Prandtl numbers need fluid properties at the proper reference temperature and geometry. In a thermally developing flow, the local Nusselt number may differ from its fully developed average. Write these distinctions directly beside the calculation rather than after the final answer.

Worked investigation: gray surface thermal radiation

We use a numerical teaching model with the following documented inputs: surface temperature = 700; Tsur = 310; epsilon = 0.8; A = 3. The primary control surface temperature is measured in K. Other parameters are Tsur [K]; epsilon [fraction]; A [m²]. Using the model equation, the predicted net outward radiation is 31.4182 kW. This is an arithmetic result for assumed properties, not a measured industrial duty or a certified equipment rating.

Increase the primary control from 700 to 772 K while holding the other inputs fixed. The same model predicts 47.0816 kW. Compare the absolute difference (15.6634 kW) with the percentage difference, and interpret the sign using the physical mechanism. Does this change chiefly alter the driving force, the thermal resistance, the material storage, or the rate at which the process moves energy? Explain why interpreting the curve without checking the other controlled inputs could be misleading.

The control range is deliberately finite. At very large or small values a simplified correlation can violate its validity conditions even if the numerical program continues to return finite numbers. For this lab, the important limitation is: Small convex diffuse gray surface facing large isothermal enclosure, view factor≈1. Decide whether the default state is inside that limitation and state what kind of additional property data, geometry measurement, or experimental coefficient would be needed to increase confidence in the estimate.

Measurement strategy and uncertainty

Derive the sign of both contributions and examine ambient temperature and air-speed uncertainty. Select one variable from the worked model that could realistically be measured on a plant floor, one that would normally be inferred, and one that is liable to drift over months of operation. Suggest how each could be independently checked. Thermal couples need installation and dynamic-response consideration, while flow meters require correct fluid properties and calibration. A sensor can be accurate at its measurement point yet irrelevant to the heat-transfer boundary being modeled.

Develop a first-order sensitivity calculation. If the output Y depends on parameter p, estimate dY/dp from two nearby runs of the simulator. If the measured parameter has uncertainty u(p), its approximate output contribution is abs(dY/dp)*u(p). This is a local linear approximation; it does not characterize poorly known geometry, phase-change instability or systematic installation bias. Keep uncertainty in the output's physical units until presenting any final percentage. Where a temperature difference is small, uncertainty can become the dominant proportion of the inferred transfer rate.

Process decision and limitation

Derive the sign of both contributions and examine ambient temperature and air-speed uncertainty. An excellent model is still a screening tool if it ignores details that control real equipment performance. A preliminary calculation may support an experiment or identify the direction of a sensitivity, but it cannot substitute for a vendor curve, verified thermophysical properties, design margins, pressure-equipment requirements or an installed sensor audit. Present the conclusion as an evidence-backed statement with a named range of validity. Avoid claiming that a temperature or duty is safe solely because one steady-state expression has a positive numerical solution.

Independent worked case

[Read a fully contextualized numerical case: Film-temperature limits](../worked-cases/unit-05-case-04.html).

Exercises

1. Starting with Qdot=A[h(Ts−Tinf)+epsilonsigma(Ts^4−Tsur^4)], identify the independent variables and produce a dimensionally consistent derivation or explanation of each term. Give an example where using the wrong unit would change the answer by at least a factor of 100.
2. Repeat the worked experiment with the control variable reduced by 10%. Compute the new output, the percent change, and compare the result to a first-order sensitivity estimate. Explain whether a linear approximation is justified.
3. Derive the sign of both contributions and examine ambient temperature and air-speed uncertainty. Describe the minimum sensor set and a method for recording uncertainty in the field.
4. Construct an alternative physical model for the same plant scenario by changing one boundary condition. State which observation could distinguish the two models.
5. Design a sensitivity chart comparing two material-property assumptions. What input dominates the uncertainty at the baseline and why?

Reproducible laboratory

Run [Laboratory 10: Gray surface thermal radiation](../labs/lab-10.html), export 151 modeled points to CSV, and annotate where the correlation is appropriate. Cite the chosen property source and write an explicit statement about the limits of extrapolation. The laboratory provides a schematic response curve and numerical readout but cannot validate a design without measured plant data.

Applied project 05: Radiation and high-temperature heating

Project brief

Thermal radiation transfers energy through electromagnetic waves. Its fourth-power temperature dependence makes it unusually important at furnace temperatures, while emissivity, geometry and participating gases limit simple blackbody calculations. Radiation is not inherently absent at low temperature; it is simply competing with other transfer paths. Build a small but defensible engineering evaluation for a hypothetical process facility. The target is not an equipment purchase specification: it is a traceable analysis of what would need to be known before design or procurement. Work with absolute temperatures and emissivity, model gray-body exchanges, and examine thermal radiation in furnaces and ovens.

Process case and data plan

The scenario begins with incomplete information. A process technician has measured bulk inlet and outlet temperatures, approximate flow and a surface temperature, while operations has an estimate of annual hours. Choose one of the four lesson topics below and define a physically plausible dataset with at least eight rows. Use SI units. You may use representative approximate fluid properties if you label them as assumptions; do not describe fabricated samples as field measurements.

• Option 1: Blackbody emission and emissivity. A hot processing tray cools to an enclosed room. Compare a polished and oxidized surface using Kelvin, and explain why Celsius to the fourth power is invalid. Begin with Qdot=epsilonsigmaA*(Ts^4−Tsur^4) and then examine which neglected effect might dominate.
• Option 2: Two-surface gray radiation. Two large furnace panels exchange heat across a narrow gap. Investigate when the large-parallel-plate approximation becomes poor because of edge view factors. Begin with Qdot=sigmaA(T1^4−T2^4)/(1/e1+1/e2−1) and then examine which neglected effect might dominate.
• Option 3: Radiation shielding and furnaces. A metal radiation shield is installed inside a furnace enclosure. State the ideal adiabatic-shield assumption and why gas convection can erode the calculated reduction. Begin with Qdot≈sigmaA(T1^4−T2^4)/Rrad and then examine which neglected effect might dominate.
• Option 4: Radiative and convective interaction. A heated pipe is exposed to moving air in a production hall. Derive the sign of both contributions and examine ambient temperature and air-speed uncertainty. Begin with Qdot=A[h(Ts−Tinf)+epsilonsigma(Ts^4−Tsur^4)] and then examine which neglected effect might dominate.

Work packages

A — Model the thermal circuit. Produce a diagram of thermal boundaries and flow directions. Record each resistance, energy-storage term, or imposed heat flux. Separate first-law calculations from empirical closure equations. Verify one result using an independent algebraic or numerical method.

B — Establish a baseline. Choose reasonable conditions, calculate heat duty in watts or kilowatts and identify one required material property that is not supplied. Define how a student would obtain that property from an authoritative source and which temperature or pressure governs the lookup.

C — Experiment and plot. Run two relevant interactive laboratories and export the curves. Compare at least three operating points, keeping all but one input fixed at a time. Include a table with the model assumption, the changed parameter and the resulting percentage change. Explain why some apparent trends would fail in a full system because stream capacity rates, pressure drop or properties co-vary.

D — Measurement and risk. Write a list of five instruments or process observations necessary to validate the simplified model. Include likely systematic and random errors, a rough range of plausible input uncertainty, and a note about unavailable information such as surface condition or contact resistance. Do not infer any pressure-vessel or worker-safety adequacy from the simulator.

E — Decision memorandum. Provide a two-page recommendation that ranks measurement, retrofit and operating-control options by estimated benefit and uncertainty. For energy economics, use explicit unit price, operating hours and equipment efficiency assumptions. Highlight which conclusions would change under a realistic alternative case.

Acceptance criteria

Deliver a readable process schematic, a table of inputs and sources, equation derivations with units, one independent verification, at least two experimental response plots, a measurement strategy and a decision memo. A technically unsupported precise number is not acceptable. The model is a teaching artifact, not a certified plant design.

Worked case 5.1: Kelvin conversion in radiation

Process situation and calculation

A steel tray at 500°C faces an enclosure at 50°C. Absolute temperatures are 773.15 K and 323.15 K. Net radiation depends on the difference 773.15^4−323.15^4, not on 500^4−50^4. The resulting fourth-power combination is physically related to thermal emission and cannot be evaluated in Celsius.

Physical interpretation and possible error

The emissivity of a rough oxidized steel tray can differ from that of polished metal. Surface condition, wavelength and geometry are often more uncertain than the constants in a calculator.

Verification tasks

1. Annotate a sketch with heat-flow directions and temperature measurement positions. Explain whether the quoted numerical value is a rate, a transport coefficient, a stored-energy scale or a transient timescale.
2. Repeat the calculation after changing one physical parameter by 10%. Identify whether the result responds linearly, logarithmically or with a power law, and derive that dependence rather than relying only on a graph.
3. State which property or correlation requires independent external verification. What uncertainty would a surface thermocouple, flow meter or geometry measurement introduce?
4. Describe the limiting condition under which the numerical relationship is not meaningful. Do not generalize beyond its stated geometry or thermal regime.

Scope: This is an illustrative, parameter-specific calculation for education, not an equipment rating or validation of actual process data.

Worked case 5.2: Parallel plate radiation denominator

Process situation and calculation

Large diffuse gray surfaces at 700 K and 350 K have emissivities 0.75 and 0.65. Their net radiative exchange per unit area is σ(700^4−350^4)/(1/0.75+1/0.65−1), approximately 6.82 kW/m². A blackbody assumption would overstate the heat exchange because it omits both surface resistances.

Physical interpretation and possible error

The large-plate formula assumes view factor nearly unity. A narrow heater panel facing a large surrounding room has a different geometrical exchange; the calculation must then use a radiative enclosure model with view factors.

Verification tasks

1. Annotate a sketch with heat-flow directions and temperature measurement positions. Explain whether the quoted numerical value is a rate, a transport coefficient, a stored-energy scale or a transient timescale.
2. Repeat the calculation after changing one physical parameter by 10%. Identify whether the result responds linearly, logarithmically or with a power law, and derive that dependence rather than relying only on a graph.
3. State which property or correlation requires independent external verification. What uncertainty would a surface thermocouple, flow meter or geometry measurement introduce?
4. Describe the limiting condition under which the numerical relationship is not meaningful. Do not generalize beyond its stated geometry or thermal regime.

Scope: This is an illustrative, parameter-specific calculation for education, not an equipment rating or validation of actual process data.

Worked case 5.3: Shield effectiveness

Process situation and calculation

Insert an ideal thin gray shield between two large parallel surfaces. Each side of the shield adds radiative surface resistance, lowering ideal exchange for the same end temperatures. With equal blackbody bounding plates and an ideal perfectly black shield, there are two series radiative gaps and the flux is half of the no-shield case.

Physical interpretation and possible error

An actual thin foil can exchange heat by convection with an intervening gas and conduction through supports. Those pathways erode the shield advantage, especially when the shield touches mounting points.

Verification tasks

1. Annotate a sketch with heat-flow directions and temperature measurement positions. Explain whether the quoted numerical value is a rate, a transport coefficient, a stored-energy scale or a transient timescale.
2. Repeat the calculation after changing one physical parameter by 10%. Identify whether the result responds linearly, logarithmically or with a power law, and derive that dependence rather than relying only on a graph.
3. State which property or correlation requires independent external verification. What uncertainty would a surface thermocouple, flow meter or geometry measurement introduce?
4. Describe the limiting condition under which the numerical relationship is not meaningful. Do not generalize beyond its stated geometry or thermal regime.

Scope: This is an illustrative, parameter-specific calculation for education, not an equipment rating or validation of actual process data.

Worked case 5.4: Film-temperature limits

Process situation and calculation

A 200°C hot pipe in 20°C air loses energy by both convection and radiation. An analyst might calculate Qconv=hA×180 K, then Qrad=εσA(Ts^4−Tsur^4) using 473.15 K and 293.15 K. If h=8 W/(m² K) and A=1 m², convection is 1.44 kW before considering radiation.

Physical interpretation and possible error

The two contributions are not strictly independent if radiation modifies nearby surface temperatures or natural convection. Iterating the wall temperature with both pathways is more realistic than declaring the convective heat rate equal to the entire thermal balance.

Verification tasks

1. Annotate a sketch with heat-flow directions and temperature measurement positions. Explain whether the quoted numerical value is a rate, a transport coefficient, a stored-energy scale or a transient timescale.
2. Repeat the calculation after changing one physical parameter by 10%. Identify whether the result responds linearly, logarithmically or with a power law, and derive that dependence rather than relying only on a graph.
3. State which property or correlation requires independent external verification. What uncertainty would a surface thermocouple, flow meter or geometry measurement introduce?
4. Describe the limiting condition under which the numerical relationship is not meaningful. Do not generalize beyond its stated geometry or thermal regime.

Scope: This is an illustrative, parameter-specific calculation for education, not an equipment rating or validation of actual process data.

Laboratory 09 — Natural convection vertical plate

Linked unit: 4. Measured output: Average h (W/(m² K)). Equation: Churchill–Chu: Nu=[0.825+0.387 Ra^(1/6)/(1+(0.492/Pr)^(9/16))^(8/27)]².

Experiment objective and theory

This interactive experiment isolates one thermal relationship so that parameter sensitivity can be observed without the distraction of a full process model. It does not solve an entire plant; the objective is to learn which inputs the mathematical expression regards as controlling, which inputs the engineer would have to obtain from measurements, and which unmodeled effects could reverse an apparent benefit. Start by writing the direction of expected heat transfer. For an entirely passive system, check that the response is compatible with the stated hot and cold temperatures and the sign convention.

Validity boundary: Ideal vertical isothermal plate; buoyancy Boussinesq β≈1/Tfilm. The simplified correlation treats the remaining conditions as fixed even if that assumption would not survive an actual change in flow, phase or material temperature. If the laboratory displays a finite number at an invalid combination, do not call that number a validated prediction. Temperature-dependent properties should be evaluated at a defensible reference state; this model deliberately does not call a property database.

Default test and sensitivity measurements

Primary variable: surface ΔT, unit K. Initial value: 40 K. Predicted output: 4.626837 W/(m² K). Your result may round on the web page. Confirm the orders of magnitude independently with the printed equation.

At the first selected test value 27.6 K, the reference model gives 4.130162 W/(m² K). At the second test value 104.4 K, it gives 6.221383 W/(m² K). Do not draw a straight-line conclusion from these two numbers without checking the full response curve. Export the CSV, calculate a central finite difference near the default value, and compare it with the curvature of the graph. Keep the parameter-change interval small enough for a sensitivity calculation but large enough to avoid rounding effects.

Controlled-variable experiment

1. Record the baseline output and plot before touching any slider. Explain whether a temperature difference, resistance, heat capacity or dimensionless correlation establishes the result.
2. Move the primary slider to its low and high limits. Identify the graph region where an extrapolation of the physics would be most dangerous.
3. Restore the baseline. Change one secondary material or geometry parameter by approximately 10%. State the expected direction before observing it; then describe the difference between your prediction and the numerical response.
4. Use reset to reproduce the default experiment, and export the CSV. Check the exported data for units, sign and monotonicity. A curve with a sharp numerical discontinuity should be investigated rather than smoothed away automatically.
5. Prepare a four-column comparison of baseline, adjusted input, adjusted output and physical interpretation. Finish with a statement of the lab model's limitations and what one real measurement would validate it.

Interpretation and field extension

A computational curve is a relationship under controlled assumptions, not a set of experimental observations. In a plant, thermal conductivity, viscosity and heat capacity change with temperature; fouling deposits grow in time; and equipment area can be effectively reduced by flow maldistribution. Explain whether the chosen experiment can distinguish these effects. If not, propose a paired measurement, such as wall temperatures and stream flow, that creates additional evidence. Never claim improved process safety or energy economics solely from a single slider result.

Questions: (a) Which input appears with the highest exponent in the local law? (b) Could an unmeasured contact resistance mimic the same trend? (c) Does the graph represent a local coefficient, an area-average value or a whole-unit thermal duty? (d) What would an alternative geometry change? (e) What independent instrument or property reference would be required before industrial scale-up?

Laboratory 10 — Gray surface thermal radiation

Linked unit: 5. Measured output: Net outward radiation (kW). Equation: Q=epsilon sigma A(Ts^4−Tsur^4).

Experiment objective and theory

This interactive experiment isolates one thermal relationship so that parameter sensitivity can be observed without the distraction of a full process model. It does not solve an entire plant; the objective is to learn which inputs the mathematical expression regards as controlling, which inputs the engineer would have to obtain from measurements, and which unmodeled effects could reverse an apparent benefit. Start by writing the direction of expected heat transfer. For an entirely passive system, check that the response is compatible with the stated hot and cold temperatures and the sign convention.

Validity boundary: Small convex diffuse gray surface facing large isothermal enclosure, view factor≈1. The simplified correlation treats the remaining conditions as fixed even if that assumption would not survive an actual change in flow, phase or material temperature. If the laboratory displays a finite number at an invalid combination, do not call that number a validated prediction. Temperature-dependent properties should be evaluated at a defensible reference state; this model deliberately does not call a property database.

Default test and sensitivity measurements

Primary variable: surface temperature, unit K. Initial value: 700 K. Predicted output: 31.41816 kW. Your result may round on the web page. Confirm the orders of magnitude independently with the printed equation.

At the first selected test value 470 K, the reference model gives 5.383898 kW. At the second test value 980 K, it gives 124.2673 kW. Do not draw a straight-line conclusion from these two numbers without checking the full response curve. Export the CSV, calculate a central finite difference near the default value, and compare it with the curvature of the graph. Keep the parameter-change interval small enough for a sensitivity calculation but large enough to avoid rounding effects.

Controlled-variable experiment

1. Record the baseline output and plot before touching any slider. Explain whether a temperature difference, resistance, heat capacity or dimensionless correlation establishes the result.
2. Move the primary slider to its low and high limits. Identify the graph region where an extrapolation of the physics would be most dangerous.
3. Restore the baseline. Change one secondary material or geometry parameter by approximately 10%. State the expected direction before observing it; then describe the difference between your prediction and the numerical response.
4. Use reset to reproduce the default experiment, and export the CSV. Check the exported data for units, sign and monotonicity. A curve with a sharp numerical discontinuity should be investigated rather than smoothed away automatically.
5. Prepare a four-column comparison of baseline, adjusted input, adjusted output and physical interpretation. Finish with a statement of the lab model's limitations and what one real measurement would validate it.

Interpretation and field extension

A computational curve is a relationship under controlled assumptions, not a set of experimental observations. In a plant, thermal conductivity, viscosity and heat capacity change with temperature; fouling deposits grow in time; and equipment area can be effectively reduced by flow maldistribution. Explain whether the chosen experiment can distinguish these effects. If not, propose a paired measurement, such as wall temperatures and stream flow, that creates additional evidence. Never claim improved process safety or energy economics solely from a single slider result.

Questions: (a) Which input appears with the highest exponent in the local law? (b) Could an unmeasured contact resistance mimic the same trend? (c) Does the graph represent a local coefficient, an area-average value or a whole-unit thermal duty? (d) What would an alternative geometry change? (e) What independent instrument or property reference would be required before industrial scale-up?

Unit 05 — Assessment: Radiation and high-temperature heating

Instructions: Work at least three problems with complete units, a diagram, assumptions and a sensitivity check. These questions emphasize independent explanation rather than recognizing a repeated numerical template.

Topic 5.1: Blackbody emission and emissivity

1. Derivation: Starting with `Qdot=epsilon*sigma*A*(Ts^4−Tsur^4)`, derive the expression from stated physical assumptions or explain the empirical closure that prevents a first-principles derivation. Define the system boundary and identify which values are measurable.

2. Engineering investigation: A hot processing tray cools to an enclosed room. Produce two competing hypotheses that could explain a 12% deviation between the calculated heat duty and the measured duty; for each give an instrument or a controlled test.

3. Critical evaluation: Compare a polished and oxidized surface using Kelvin, and explain why Celsius to the fourth power is invalid. State how the result would change for a tenfold increase in one relevant physical length, and whether the original correlation remains valid.

Topic 5.2: Two-surface gray radiation

1. Derivation: Starting with `Qdot=sigma*A*(T1^4−T2^4)/(1/e1+1/e2−1)`, derive the expression from stated physical assumptions or explain the empirical closure that prevents a first-principles derivation. Define the system boundary and identify which values are measurable.

2. Engineering investigation: Two large furnace panels exchange heat across a narrow gap. Produce two competing hypotheses that could explain a 12% deviation between the calculated heat duty and the measured duty; for each give an instrument or a controlled test.

3. Critical evaluation: Investigate when the large-parallel-plate approximation becomes poor because of edge view factors. State how the result would change for a tenfold increase in one relevant physical length, and whether the original correlation remains valid.

Topic 5.3: Radiation shielding and furnaces

1. Derivation: Starting with `Qdot≈sigma*A*(T1^4−T2^4)/Rrad`, derive the expression from stated physical assumptions or explain the empirical closure that prevents a first-principles derivation. Define the system boundary and identify which values are measurable.

2. Engineering investigation: A metal radiation shield is installed inside a furnace enclosure. Produce two competing hypotheses that could explain a 12% deviation between the calculated heat duty and the measured duty; for each give an instrument or a controlled test.

3. Critical evaluation: State the ideal adiabatic-shield assumption and why gas convection can erode the calculated reduction. State how the result would change for a tenfold increase in one relevant physical length, and whether the original correlation remains valid.

Topic 5.4: Radiative and convective interaction

1. Derivation: Starting with `Qdot=A[h(Ts−Tinf)+epsilon*sigma*(Ts^4−Tsur^4)]`, derive the expression from stated physical assumptions or explain the empirical closure that prevents a first-principles derivation. Define the system boundary and identify which values are measurable.

2. Engineering investigation: A heated pipe is exposed to moving air in a production hall. Produce two competing hypotheses that could explain a 12% deviation between the calculated heat duty and the measured duty; for each give an instrument or a controlled test.

3. Critical evaluation: Derive the sign of both contributions and examine ambient temperature and air-speed uncertainty. State how the result would change for a tenfold increase in one relevant physical length, and whether the original correlation remains valid.

Assessment rubric

Conceptual accuracy (30%): state the correct physics, not just a numerical recipe. Traceable units and sign conventions (20%): make every conversion explicit. Sensitivity and validity (25%): quantify a change and identify non-applicable regimes. Engineering communication (25%): diagrams, model limits and measurements must support the recommendation.

The exercises are deliberately open; different property assumptions can produce different numerical outcomes. Clearly distinguish supplied values, literature property values and measurements. A plausible order of magnitude with transparent limitations is more useful than an unjustifiably precise number.