0
$\begingroup$

I am doing an experiment regarding heat transfer. I put aluminium pieces of same mass (about 15.3g) but different exposed surface area. To create this, I cut 4 aluminium pieces into same dimension (16cm x 4cm), then folded them into dimension of (12cm x 4cm), (8cm x 4cm) and (4cm x 4cm). Then I put the pieces in the oven set to 150 ºC. After the temperature is reached, I took the aluminium and dunk it in water (25ºC and about 0.125 kg) while recording the temperature of the water through a temperature sensor.

The results is that we should expect a exponential decay curve from solving Newton's law of cooling.

For the 8cm x 4cm piece (Temperature(ºC) vs time(s)): Temperature per time curve 8x4 (Theres a tiny bump shown in the circle but I assumed its uncertainty due to sensor and environmental factors so I considered it as an exponential decay, is it actually uncertainty or related to what explained below?)

This is close to the exponential decay.

But for 4cm x 4cm the (Temperature (ºC) vs time (s)) curve is weird (not really exponential decay).

Temperature per time for 4x4

There is a noticeable decrease in the temperature of the water after it reach max temperature (a bit before 20 seconds). I thought it may be due to evaporation but I was not sure. What is happening? And at what point do I considered that the heat finish transferred?

Edit 1: Reply to comments on needing relative position of the temperature sensor. Note: the water covers all of the aluminium and the temperature sensor. Picture of experimental setup

$\endgroup$
3
  • $\begingroup$ Where is your sensor relative to the location of the aluminum pieces? $\endgroup$
    – Bob D
    Commented Sep 5, 2019 at 2:19
  • $\begingroup$ I'd like to answer this question tomorrow, it's very late here. Unless someone beats me to it, of course. Questions with empirical data should always be honoured here, I strongly believe. $\endgroup$
    – Gert
    Commented Sep 5, 2019 at 2:54
  • $\begingroup$ It is necessary to submit a photo of the samples, their position in the water, the position of the sensor. $\endgroup$ Commented Sep 5, 2019 at 4:00

1 Answer 1

0
$\begingroup$

It's always great to see a contributor actually carrying out some experimentation, so welcome!

However, I feel your experimental protocol leaves somewhat to be desired to.

I took the aluminium and dunk it in water (25ºC and about 0.125 kg) while recording the temperature of the water through a temperature sensor.

150ºC is higher than the BP of water. It will cause temporary, localised boiling and you don't want that. Don't exceed 100ºC as initial temperature of the $\text{Al}$.

Your pieces of $\text{Al}$ seem very close together, this may reduce the effective surface area available for convection. Spread them out more.

Your set-up isn't insulated at all and the shallow water promotes heat loss to the environment. Use something like a styrofoam coffee cup (with lid) to avoid losses.

It may also be simpler and more accurate to measure the temperature increase of the water after some fixed period of time, $\Delta t$.

(Theres a tiny bump shown in the circle but I assumed its uncertainty due to sensor and environmental factors so I considered it as an exponential decay, is it actually uncertainty or related to what explained below?)

You're assuming exponential evolution of the water temperature because you're thinking of Newton's Cooling Law (I assume). Because here both the water and $\text{Al}$ temperatures change, it's cautious to verify that, as I'll do below.


All parameters of $\text{Al}$ are indexed as $1$, all parameters of water are indexed as $2$.

Newton's Cooling Law teaches us that:

$$\frac{\text{d}Q}{\text{d}t}=-hA(T_1-T_2)\tag{1}$$ Assuming no heat losses (adiabatic system), then: $$\text{d}Q=m_1c_1\text{d}T_1=-m_2c_2\text{d}T_2$$ $$m_1c_1\frac{\text{d}T_1}{\text{d}t}=-hA(T_1-T_2)\tag{2}$$ $$-m_2c_2\frac{\text{d}T_2}{\text{d}t}=-hA(T_1-T_2)\tag{3}$$ Divide $(2)$ by $(3)$: $$-\frac{m_1c_1\text{d}T_1}{m_2c_2\text{d}T_2}=1$$ $$\text{d}T_1=-\alpha \text{d}T_2\tag{4}$$ Where: $$\alpha=\frac{m_2c_2}{m_1c_1}$$ Integrate $(4)$: $$\int\text{d}T_1=\int-\alpha \text{d}T_2\to T_1=-\alpha T_2+C_1$$ Determine integration constant $C_1$ from initial conditions: $$T_{1,0}=-\alpha T_{2,0}+C_1$$ $$C_1=T_{1,0}+\alpha T_{2,0}$$ Insert all this into $(3)$: $$-m_2c_2\text{d}T_2=-hA(-\alpha T_2+C_1-T_2)\text{d}t$$ $$m_2c_2\text{d}T_2=hA[C_1-(\alpha +1)T_2]\text{d}t\tag{5}$$ Make a simple substitution: $$u=C_1-(\alpha +1)T_2$$ So that: $$\text{d}u=-(\alpha +1)\text{d}T_2\to \text{d}T_2=-\frac{1}{\alpha +1}\text{d}u$$ Insert into $(5)$: $$-\frac{1}{\alpha +1}\frac{{d}u}{u}=hA\text{d}t$$ $$-\frac{1}{\alpha +1}\int_{u_0}^{u(t)}\frac{{d}u}{u}=\int_0^t hA\text{d}t$$ $$\ln\frac{u(t)}{u_0}=-(\alpha +1)hAt$$ $$\frac{u(t)}{u_0}=\exp[-(\alpha +1)hAt]$$ $$\frac{C_1-(\alpha +1)T_2(t)}{C_1-(\alpha +1)T_{2,0}}=\exp[-(\alpha +1)hAt]$$ $$\boxed{T_2(t)=\frac{1}{\alpha +1}\Big[C_1-(\alpha+1)T_{2,0}\exp[-(\alpha +1)hAt]\Big]}$$

So that clearly the water temperature $T_2$ evolves at an exponential rate. At $t=+\infty$ both temperatures of $\text{Al}$ and water will have become equal and invariant. These final temperatures will be independent of both $h$ and $A$.

$$t=+\infty \to C_1-(\alpha +1)T_{2,\infty}=0$$ $$T_{1,0}+\alpha T_{2,0}-(\alpha +1)T_{2,\infty}=0$$ $$T_{2,\infty}=\frac{T_{1,0}+\alpha T_{2,0}}{\alpha +1}$$

We can also determine the cooling rate:

$$\boxed{\frac{\text{d}T_2(t)}{\text{d}t}=-(\alpha+1)T_{2,0}hA\exp[-(\alpha +1)hAt]}$$

Clearly, the cooling rate is proportional to $A$.

$\endgroup$
4
  • $\begingroup$ Thank you for the suggestion and the proof. Sorry for my bad English (I have a bit of trouble understanding some part), are you suggesting that the weird temperature curve is due to the localised boiling? And also I am looking at the time at which the heat transfer is completed (or approximately completed). $\endgroup$
    – pedpx2
    Commented Sep 5, 2019 at 15:25
  • $\begingroup$ I don't know what caused the blip but localised boiling may be it. In theory the heat transfer is completed after $t=+\infty$ and is independent from both $A$ and $h$. So it's better to measure water temperature after a short while, when heat transfer is still in progress. $\endgroup$
    – Gert
    Commented Sep 5, 2019 at 15:29
  • $\begingroup$ But why does the weird curve only occur to the 4cm x 4cm piece but the other pieces (8cm x 4cm and 12cm x 4cm) does not have the weird curve and have the characteristic of exponential decay as you've suggested. $\endgroup$
    – pedpx2
    Commented Sep 6, 2019 at 9:54
  • $\begingroup$ If you suspect a data point is faulty the rule in experimentation is to repeat that specific experiment. Flukes do occur and can be very difficult to explain rationally. $\endgroup$
    – Gert
    Commented Sep 6, 2019 at 13:33

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.