Which ai to use for drafting exam and quiz questions

Software
,
1

Any tools or advice on which ai tool(s) to use for help in drafting quizzes and exams?

2

Depending on the topic you’ll get pretty good results for quizzes on all of the chatbot tools (i.e., for fun).

I would be careful with exams as you need to be absolutely sure the answers are accurate and complete, which any AI tool is not guaranteed to be able to do.

I recommend using a prompt formula to ensure that the answers are pitched at the audience you are targeting (the questions are different if you are doing a pub quiz vs. a PhD pop quiz).

This is a useful introduction to prompting: https://www.youtube.com/watch?v=jC4v5AS4RIM

1 Like
3

Here’s an example:

You are a undergrad lecturer setting a quiz about thermodynamics. Create a quiz containing 10 questions that will test the key concepts of thermodynamics and use exam-style questions requiring approximately 2-3 sentences. Here is the curriculum:

I then copy and pasted the Course Description on this website to make the questions realistic and relevant to what I am “teaching”: Course Catalogue - Thermodynamics 3 (MECE09010)

I got 10 questions back that relate back to the key areas.

I strongly recommend you know the topic well enough to answer the questions yourself - ChatGPT and other tools shouldn’t be used as accurate answers.

For those interested, here are the questions:

Thermodynamics Quiz
Topic 1: Revision

Question 1:
Explain how the First Law of Thermodynamics is applied to closed systems and provide the equation that represents this law for a closed system undergoing a cycle.

Question 2:
Describe what is meant by an isentropic process and provide an example of a scenario where this might ideally occur in a thermodynamic system.

Question 3:
Compare the thermal efficiency of a Carnot cycle to any real-world cycle. Why can’t real-world cycles achieve the efficiency of the Carnot cycle?

Topic 2: First Law Applied to Open (Flow) Systems

Question 4:
In a Brayton cycle, why is the thermal efficiency lower than that of a Carnot cycle operating between the same two temperature limits?

Question 5:
Describe how interstage cooling in a compressor can lead to a reduction in the work of compression. Explain using the concept of compressor work and ideal gas behavior.

Question 6:
What is the regenerator effectiveness in a regenerative cycle, and why is there a limitation to the efficiency that can be achieved by regenerative cycles?

Topic 3: Gas Mixtures

Question 7:
Using Dalton’s and Amagat’s laws, explain how the thermodynamic properties of an ideal gas mixture can be determined.

Question 8:
Define dew point and relative humidity. How are they relevant in the context of gas-vapour mixtures?

Topic 4: Steam and Combined Cycles

Question 9:
Compare the ideal Rankine cycle to the Carnot cycle. What are the main reasons that the efficiency of a real Rankine cycle is less than that of the Carnot cycle?

Question 10:
Discuss one method by which the efficiency of the Rankine cycle can be increased, and explain why this method is effective.

3 Likes
4

Agree re: promptcrafting.

I like giving it some text and asking it to generate questions about the text. I usually ask for multiple redundant questions so that I can pick the best one(s), and then modify them even more myself.

E.g., “Below is a description of entrepreneurship. Create three multiple choice questions based on the description.”

2 Likes
5

Oh boy! This is too good to pass up.

  1. Closed systems are used as the beginning point for the first law, ostensibly to allow an easier appreciation of conservation of energy. Nothing restricts applying the first law to an open system with appropriate expansion of the formulation. Indeed, the integral form of the first combined law contains enthalpy (with mass) flows through boundary surfaces, changes in the control volume with time, and changes in heat and work. The expression for a cycle is simply Delta U = 0.

  2. Isentropic means “constant entropy”. A practical case is the entropy change in a system undergoing a perfectly adiabatic process. No heat flow in/out means no entropy change internally.

  3. Carnot is the best case for any cyclic process between two temperatures (heat engines, heat pumps, or refrigerators). In first principles, the process is modeled to occur for ideal gases. In practice, the process has no friction losses (no entropy loss to the surroundings). Real world systems are never ideal gases and always have entropy loss to the surroundings.

  4. … Ugh!! Brayton … must be mechanical engineering … ugh ugh ugh … (why not just stick to Carnot) … ugh!

  5. Cooling removes heat. For an ideal gas, lowering temperature reduces volume at constant pressure. Alternatively, it reduces pressure at constant volume. Either of these two are alternatives to Delta(pV) = work.

  6. … limits are set by Carnot and real-world efficiency ???

  7. Dalton says, each component in a mixture behaves with its own “partial” pressure, and we sum partial pressures to get total. Amagat says, each component in a mixture behaves with its own “partial” volume, and we sum partial volumes to get total.

  8. Dew point is the temperature where a gas mixture forms its first drop of liquid at a constant pressure. Relative humidity is the percentage of water vapor in a gas relative to the amount of water vapor that could exist if the gas at that temperature was at its dew point.

  9. Rankine is less because the gas is real (it condenses). Ideal gases (Carnot) never condense.

  10. … run at higher temperature … ???

I imagine as a ChemE that I’ve obtained just an above average on what is clearly a MechE quiz. :slight_smile:

Here are a few back at you …

  1. What three core assumptions define an ideal gas versus a real gas? Which two are the first ones relaxed to create real gas equations of state?
  2. What is fugacity? Explain it in terms of fundamentals for molecules.
  3. Air is often presumed to be a pure gas. Under what practical conditions does this assumption fail?

ps – IMHO, this is one fairly intense “quiz”, both for the students to tackle and for your own post grading efforts.

–
JJW

2 Likes
6

Lots of good advice here:

Ethan Mollick is writing very interesting stuff about the LLMs available now.

7

“Official” answers:

  1. The First Law of Thermodynamics for closed systems undergoing a cycle states that the net heat transfer is equal to the net work done. The equation representing this is ( Q_{\text{net}} = W_{\text{net}} ), where ( Q ) is heat and ( W ) is work.

  2. An isentropic process is a reversible adiabatic process, meaning no heat is transferred to or from the system, and entropy remains constant. This can ideally occur in turbines and compressors where, assuming no friction or other inefficiencies, the fluid expands or is compressed without heat exchange.

  3. The thermal efficiency of a Carnot cycle is the maximum possible efficiency between two temperature limits and is not achievable in real-world cycles due to factors like friction, heat losses, and material limitations that introduce irreversibilities and reduce efficiency.

  4. The thermal efficiency of a Brayton cycle is lower than that of a Carnot cycle due to irreversibilities in the compression and expansion processes and the fact that heat addition and rejection in the Brayton cycle are not isothermal as in the Carnot cycle.

  5. Interstage cooling in compressors reduces the work required for compression because cooling the gas between stages lowers its volume, which reduces the work needed to compress it further according to the ideal gas law.

  6. Regenerator effectiveness is a measure of how much heat is transferred from the turbine exhaust to the compressed air. The efficiency gain from regenerative cycles is limited by the second law of thermodynamics, as complete heat transfer cannot be achieved due to the finite temperature difference.

  7. Dalton’s Law states that the total pressure of a gas mixture is the sum of the partial pressures of individual gases, while Amagat’s Law suggests that the total volume is the sum of the partial volumes. These laws are used to determine the properties of ideal gas mixtures by relating them to the properties of the individual gases.

  8. Dew point is the temperature at which air becomes saturated with vapor and condensation begins. Relative humidity is the ratio of the partial pressure of water vapor to the equilibrium vapor pressure at the same temperature. Both are critical for designing systems involving humidity control and predicting condensation.

  9. The ideal Rankine cycle differs from the Carnot cycle as it involves the practical limits of phase change of the working fluid, typically water. The efficiency of a real Rankine cycle is less than that of the Carnot cycle due to irreversibilities in the actual components, such as turbines and pumps.

  10. One method to increase the efficiency of the Rankine cycle is to superheat the steam, which lowers the moisture content at the end of the expansion process, reducing turbine blade erosion and increasing the work output per unit mass of steam, thus improving efficiency.

8

  1. The three core assumptions defining an ideal gas are: (1) the gas molecules themselves occupy no volume, (2) there are no intermolecular forces between the gas molecules, and (3) all collisions between gas molecules and between molecules and the walls of the container are perfectly elastic. The first two assumptions are relaxed for real gas equations of state, acknowledging that gas molecules have finite volume and experience intermolecular forces.

  2. Fugacity is a thermodynamic property that corrects the pressure of a real gas to account for non-ideal behavior at a molecular level. It is essentially an effective pressure that represents the escaping tendency of a molecule from the phase and is used to estimate the ‘chemical potential’ of a real gas, aligning with the ideal gas behavior under standard conditions.

  3. The assumption that air is a pure gas fails under practical conditions where the composition of air changes significantly, such as at high altitudes where the partial pressures of gases differ from sea level, or in industrial environments where air may be mixed with pollutants or other gases, altering its properties.

9

I actually know somewhere between nothing and very little about this topic! It was the first “clever” topic that came to mind!

10

I use GPT-4 (my university pays for it luckily). I get great results when I do some prompt engineering to get it right.

Even the students get great practice questions with GPT-3.5 as not all of them subscribe to get access to GPT-4 - I encourage them to use it for practice. I teach JavaScript and PySpark so it may not be the case with all subjects.

It saves me a bunch of time for sure.

11

Wow. That’s a significant signal of the permanence of these tools in academia.

(Meanwhile I keep hearing about how we should set up TurnItIn etc … sigh)