Heating Time Calculation: Formulas, Worked Examples and Losses

To calculate heating time, divide the energy the load needs by the power that actually reaches it: t = m·c·ΔT / (P·η). That gives a floor, the fastest the job can happen. Whether it is also a good estimate depends on two things: how much of the heater’s power leaks to the surroundings, and whether the object heats evenly or its core lags behind its surface. This guide walks through the formula, a worked example, the loss terms, and the checks that tell you when to move from a hand calculation to a transient model.
TL;DR
Energy needed is Q = m·c·ΔT. Time is Q divided by heater power times efficiency: t = Q / (P·η).
Losses matter. A constant loss subtracts from heater power, and a loss that grows with temperature makes heating slow down near the target.
Check the Biot number (Bi = hL/k). Below about 0.1 the whole object can be treated as one temperature; above it, the core lags and the simple answer is optimistic.
Phase changes (melting, boiling) need a latent-heat term, not just specific heat.
Our Excel-based toolkit combines conduction, convection and radiation with a built-in property database, so you can move past the single-number estimate.
Table of Contents
The basic equation
Heating something takes energy. For a temperature change with no phase change, the sensible heat is:
Q = m · c · ΔT
where m is mass (kg), c is specific heat (J/kg·K) and ΔT is the difference between target and starting temperature. Textbooks such as Cengel’s Heat Transfer: A Practical Approach build heating problems on exactly this energy balance.
Time follows from power. If a heater delivers P watts (joules per second) and a fraction η of that reaches the load:
t = Q / (P · η)
Two habits keep this honest:
Work in SI units throughout. Joules, kilograms, kelvin (or °C differences), watts. Mixing Btu, lb and °F is where most arithmetic slips happen.
Count each loss once. If you fold losses into η, do not also subtract them from P. Double counting makes the estimate too slow, and it is easy to do when you reuse an efficiency figure from a datasheet.
The specific heat should be the value at the average temperature of the process, not at room temperature, when the range is wide. Cengel’s steel furnace-style examples and the furnace method discussed below both pick a mid-range value for that reason.
Worked example: water in a teapot
Cengel’s Example 1-2 is a clean reference case. Heat 1.2 kg of water from 15°C to 95°C in a teapot with a 1200 W element. The teapot itself is 0.5 kg with c = 0.7 kJ/kg·°C, water is 4.18 kJ/kg·°C, and heat loss is neglected.
Water: 1.2 × 4.18 × 80 = 401.3 kJ
Teapot: 0.5 × 0.7 × 80 = 28 kJ
Total: about 429.3 kJ
Time: 429.3 kJ ÷ 1.2 kJ/s ≈ 358 s, about 6.0 minutes
The book notes that the real time is longer, because heat loss is unavoidable. That one sentence is the whole story of this topic: the calculation above is a lower bound.
Notice also that the container was part of the load. Forgetting the 0.5 kg teapot would have understated the energy by about 7%. For a heavy steel vessel holding a little liquid, the container can dominate.
A second reference case, TU Delft’s energy balance examples, asks how long a 1 kW cooker needs to heat 0.5 L of water from 20°C to 80°C, with water at 4200 J/kg·K and no losses or cooker heat capacity. The energy is 0.5 × 4200 × 60 = 126 kJ, so the time is 126 s. The same course then repeats the problem with a heat-loss term, which is the next step.
Adding heater efficiency and heat losses
There are two practical ways to include losses.
Option 1: a single efficiency factor
Take the TU Delft case at 90% efficiency: t = 126 kJ / (1 kW × 0.9) = 140 s. The answer is quick and fine for early sizing. Its weakness is that efficiency is not constant. Losses start near zero when the load is cold and climb as it gets hot, so an efficiency figure is an average that hides slower heating near the target.
For furnaces, Thermal Processing Magazine suggests a way to estimate that figure: divide the power needed to hold an empty, temperature-soaked furnace at the target temperature into the furnace’s full power. It notes that well-insulated furnaces are roughly 80 percent efficient or better, while poorly insulated ones are much worse.
Option 2: a loss term in the energy balance
If the loss is a known constant, subtract it from the power. If it depends on temperature, model it. TU Delft’s second example uses a linear loss, q = β(T − T∞), with β = 5.3 W/K and T∞ = 20°C, and concludes the time must exceed the no-loss answer. With the loss growing as the water warms, you step through time or solve the resulting first-order equation.
The Watlow power calculation guide lays out the full heater-sizing method:
Sensible heat: Q = w·Cp·ΔT, with latent terms (w·Hf for melting, w·Hv for vaporizing) added when the material changes phase.
Conduction loss through insulation: QL1 = k·A·ΔT·te / (3.412·L).
Convection loss with orientation factors of 1.29 for a surface facing up, 1.00 vertical and 0.63 facing down.
Radiation loss: QL3 = A·FSL·e.
For start-up, use two thirds of the full heat loss, because losses rise from zero to 100% during heat-up, then add a safety factor.
Watlow also points out that losses can be a small fraction of the total power or the largest portion of it, so they should not be ignored unless experience says it is safe.
For wall and enclosure losses specifically, our guide on how to calculate heat loss covers the conduction, convection and radiation paths. If you are choosing the element itself, see heater coil sizing.

The power-versus-loss check
Before computing any time, compare heater power with the loss at the target temperature. If the loss at the final temperature is equal to or larger than the heater’s output, the object never reaches the target, however long you wait. Every heating-time formula quietly assumes the heater has headroom. A quick loss estimate at the final temperature costs one line and prevents an impossible answer.
When energy over power is enough
The formula t = Q/(P·η) treats the object as one temperature. That is reasonable when:
the heater power is the limit (a strong heat source is not the case; a modest heater on a thermally small load is),
the material conducts heat well relative to how fast heat arrives at its surface, and
there is no phase change in the range.
A more precise test comes from transient conduction theory. In the lumped capacitance method, covered in NTU’s open course notes on time-dependent conduction, an energy balance over the solid gives
(T − T∞)/(Ti − T∞) = exp(−h·As·t / (ρ·V·c))
The group τ = ρ·V·c / (h·As) is the thermal time constant, and the result is valid when internal temperature gradients are small.
The Biot number decides
The Biot number, Bi = h·Lc/k with Lc = V/As, compares the resistance to heat flow inside the object with the resistance at its surface. The usual threshold is Bi ≤ 0.1. Below it, one temperature describes the object. Above it, the surface is hot while the core is still cold, and the lumped time underestimates how long the core takes. That is the unsafe direction for a heating time. Our posts on the Biot number criterion and Bi and Fourier number limits for transient conduction cover the check in detail, and lumped capacitance with two worked examples shows it in use.
Heating toward a medium: the exponential form
When an object is heated by a hot fluid or furnace at temperature T∞ (rather than by a fixed-power element), the temperature approaches T∞ exponentially:
T(t) = T∞ + (Ti − T∞)·e^(−t/τ)
Solving for the time to reach a target:
t = −τ · ln[(Ttarget − T∞) / (Ti − T∞)]

Two consequences follow. After one time constant the object has closed about 63% of the gap, after three about 95%. And a target equal to or above T∞ is unreachable, since the log argument becomes zero or negative. This is why the furnace article above advises starting the soak clock when the part center is within 10 to 25°F (6 to 14°C) of set point: the center creeps toward the set point asymptotically, and waiting for exact equality would never end.
A quick decision table
Situation | Use | Watch for |
Fixed-power heater, small object, low Bi | t = m·c·ΔT / (P·η) | Container mass, realistic η |
Same, but large or poorly insulated vessel | Energy balance with loss term, stepped in time | Loss at target exceeding heater power |
Heated by a hot fluid or furnace, Bi ≤ 0.1 | Lumped exponential, t = −τ ln(…) | h varying with temperature |
Thick or low-conductivity part, Bi > 0.1 | Transient conduction (Fo, charts, numerical) | Core lag, soak criteria |
Melting or boiling in range | Add latent heat terms | Sensible-only formula fails |
Specific cases worth knowing
Phase change
The sensible-heat formula does not apply across melting or boiling. TU Delft’s third example asks how long the same 1 kW cooker takes to fully evaporate 0.5 L of water already at 100°C, which needs a mass balance plus a latent-heat balance. Add the latent term (mass times latent heat) to the sensible terms before dividing by power.
Air
Air follows the same energy logic but behaves differently in practice. Cengel treats air as an ideal gas with constant properties, using about 0.240 Btu/lbm·°F to heat a room’s air from 50°F to 70°F, and 1.007 kJ/kg·°C at the average temperature for a duct. Air carries far less energy per unit volume than water because of its low density and specific heat, so the energy figure looks small. The catch is the surroundings: the walls, furniture and structure are the real thermal mass, and in ducts the mass flow rate sets the energy rate. A room heating time based on air alone is badly optimistic.
Large tanks
For a tank, the exponential balance with τ = C/UA (thermal capacitance over the loss conductance) is the right frame. A well-insulated tank may need little power just to hold temperature, while a bare tank with a small heater might struggle to get there at all. That brings you back to the power-versus-loss check.
Cooling is the mirror image
Cooling follows the same exponential logic with the sign reversed, so a cooldown estimate uses the same time constant. See our cooldown time calculation guide for the formulas and pitfalls.
A repeatable workflow
Define the load: mass, material, container, start and target temperatures.
Pick specific heat at the average temperature. Include the container.
Compute Q, adding latent terms for any phase change.
Estimate loss at the target temperature and confirm the heater exceeds it.
Compute the ideal time Q/P, then apply a realistic efficiency or a loss term, not both.
Check Bi. If it is above 0.1, treat the result as optimistic and move to a transient conduction model.
Validate against a measurement. A thermocouple at the center of the slowest part gives the real heating curve, which you can use to calibrate η or h.
Perspective
Most heating-time errors we see in the engineering literature are not arithmetic errors. They come from using one number, usually efficiency, to stand in for physics that changes during the heat-up. Our view is that the hand formula is where you start and a model with real conduction, convection and radiation is where you finish. The reason we built the Jewlz Engineering Toolkit as one Excel-based workbook is so you do not have to jump between tools: material properties that vary with temperature come from a built-in database, and the three heat transfer modes sit in the same place across a wide temperature range. We also offer free engineering software, because students and small teams should be able to check a first estimate without a licensing hurdle. Joel
Take the estimate further
When the hand calculation tells you Bi is above 0.1, or losses depend strongly on temperature, you need a transient model rather than a single efficiency number. The Jewlz Engineering Toolkit is a downloadable Excel-based tool for conduction, convection and radiation analysis, with variable material properties and a built-in property database. If your heating problem involves a fluid, our thermal analysis tools and CFD simulation pages show what is available.

For heat exchangers, where heating time and duty are linked, see heat exchanger optimization. For the wider modeling picture, our heat transfer model guide covers methods and math, and the heat conduction calculation post covers the conduction side.
FAQ
How do I calculate how long it takes to heat water?
Compute the energy with Q = m·c·ΔT, using about 4180 to 4200 J/kg·K for water, then divide by the heater power times efficiency. For 2 kg of water from 10°C to 100°C on a 1.2 kW kettle at 90% efficiency, that is 4180 × 2 × 90 / (0.9 × 1200), about 697 s or roughly 12 minutes. Include the container mass if it is significant.
Why is my real heating time longer than the calculation?
The formula gives the minimum time with all power going into the load. Heat leaks to the surroundings, and the leak grows as the load gets hotter, so heating slows near the target. Specific heat also changes with temperature, and any phase change needs extra energy. Apply a loss term or a realistic efficiency, such as 80% or better for a well-insulated furnace.
When can I treat an object as one uniform temperature?
When the Biot number, h·(V/A)/k, is at or below about 0.1. Then internal conduction is fast compared with surface heat transfer, and the lumped capacitance method gives a reliable time. Above that, the core lags the surface and the lumped time is too short, so use a transient conduction method.
Does the same formula work for heating air?
The energy balance is the same, with air’s specific heat of about 1.007 kJ/kg·°C. But air holds little energy per volume, and in a room the walls and contents dominate the thermal mass. In ducts, use mass flow rate in the energy balance.
Sources
Recommended


Comments