Warrant¶
Comprehensive Worked Example¶
With Scaffolding Annotations and Evaluation Notes¶
How to Use This Document¶
This document contains one complete Warrant entry, written for a flash separation simulation using an ethanol-water system. The simulation context is not drawn from any homework assignment in this course - it uses familiar block types (SEP, FLASH2, HEATER) applied to a different chemical system. This is deliberate: the scaffolding annotations explain the thinking process behind each field, not the content of any specific homework.
For Students¶
Read this document once before writing your first Warrant entry. Focus on the scaffolding annotations - the amber boxes - which explain how the student arrived at each field, not just what they wrote. The entry itself is a strong example of what specificity looks like. Read it, then close this document and write your own entry from your own simulation. Do not return to this document while writing.
For Instructors¶
The rubric calibration notes at the end of this document are the primary tool for developing and iterating the Warrant rubric over time. The strong and weak versions of each field, read alongside the calibration notes, define the boundaries between rubric levels. As you read student entries across the semester, return to this document to check your calibration - particularly for Section 1.0, where the boundary between a genuine belief update and a content summary is the most difficult judgment call.
Simulation Context¶
Ethanol-water flash separation. A feed stream of 100 lbmol/hr containing 30 mol% ethanol and 70 mol% water at 80 F and 14.7 psia enters a SEP block that routes 95% of the ethanol to one product stream. The remaining stream - mostly water with trace ethanol - enters a HEATER block that raises the temperature to 200 F at 14.7 psia before entering a FLASH2 unit operating at 14.7 psia with a duty specification of zero (adiabatic). Property method: IDEAL.
This simulation was chosen for the worked example because:
-
It uses block types introduced early in the course - SEP, HEATER, FLASH2
-
The IDEAL property method creates a genuine and detectable problem for a hydrogen-bonding system
-
The SEP versus FLASH2 distinction is the central conceptual shift the entry demonstrates
-
The ethanol-water system is familiar enough to be legible without Aspen knowledge
Before the Simulation¶
The following questions are given to students before they open Aspen. They are not graded. Their purpose is to prime the student to notice things worth recording - so that the simulation generates raw observations rather than just output.
| PRE-SIMULATION PROMPT CARD |
Answer in one sentence each. Do this before opening Aspen.
|
| SCAFFOLDING ANNOTATION |
These four questions map directly to the four Warrant fields. Q1 primes Section 1.0 - the student who has written a one-sentence prediction before running the simulation has something specific to compare against the output. Q2 primes Section 2.0 - the SEP versus FLASH2 distinction. Q3 and Q4 prime Fields 02 and 04 - the IDEAL assumption and its failure condition. The most important question is Q4. A student who can answer 'no' before running the simulation - who already knows that convergence is not validation - is positioned to write a genuine Section 4.0 rather than a generic one. A student who answers 'yes' will have their assumption tested by the simulation and may produce a stronger Section 1.0 as a result. Release this card at the start of the session. Collect the written answers on paper before students open Aspen. Do not grade them - but do read them quickly as students work. The answers tell you what the class expected before running. That gap between expectation and output is where the Warrant entry lives. |
The Entry - Strong Version (Section-Based)¶
Section 1.0 - Prediction¶
| 1.0 - Prediction CONSTRAINT |
| I assumed the SEP block and the FLASH2 block were both separating streams - different tools for the same job. After running the simulation and looking at both outputs, I no longer believe that. The SEP block produced exactly the split I specified: 95% of the ethanol to one stream, regardless of whether that split is thermodynamically achievable at these conditions. The FLASH2 block produced a split I did not specify - it calculated what vapor and liquid phases form at 200 F and 14.7 psia based on the equilibrium thermodynamics of the system. These are fundamentally different operations. The SEP result can be physically impossible and the simulation will not tell me. The FLASH2 result is constrained by thermodynamics and will fail to converge if I ask for something the system cannot do. |
| SCAFFOLDING ANNOTATION |
HOW THE STUDENT ARRIVED HERE: The pre-simulation prompt card asked which block enforces a split the student chose versus which calculates one from thermodynamics. A student who answered that question before running the simulation arrived at the output with a specific prediction to test. When the SEP block produced exactly 95% ethanol split regardless of conditions, and the FLASH2 block produced a vapor fraction the student did not specify, the contrast was visible. WHAT MAKES THIS STRONG: The belief update names two specific things - what the student believed before (both blocks separate streams, different tools for the same job) and what changed it (SEP enforces, FLASH2 calculates, and these are fundamentally different). The last sentence is the strongest part: 'The SEP result can be physically impossible and the simulation will not tell me.' That sentence demonstrates the student understands the epistemological consequence of using SEP, not just its operational difference from FLASH2. WHAT TO WATCH FOR IN STUDENT ENTRIES: A weak Section 1.0 summarizes content - 'I learned about the difference between SEP and FLASH2.' A strong Section 1.0 records a specific belief that changed and names what changed it. The test: can you identify what the student believed before, and what specific observation changed it? |
Section 2.0 - Simulation Setup¶
| 2.0 - Simulation Setup FRAGILITY |
| I assumed IDEAL is an acceptable property method for ethanol-water because the assignment specified it and the simulation converged without errors. This assumption fails for any calculation where the non-ideal mixing behavior of ethanol and water matters - specifically, phase equilibrium predictions near the known azeotrope at approximately 89 mol% ethanol at atmospheric pressure. IDEAL assumes activity coefficients of 1.0 for both components, which means it predicts vapor-liquid equilibrium as if the molecules had no preference for their neighbors. Ethanol and water interact strongly through hydrogen bonding - their activity coefficients are measurably greater than 1.0 across the full composition range. A FLASH2 simulation using IDEAL near the azeotrope composition would predict the wrong vapor fraction and the wrong product compositions. The simulation converged, but convergence confirmed that the calculation completed - not that the property method is appropriate. |
| SCAFFOLDING ANNOTATION |
HOW THE STUDENT ARRIVED HERE: The pre-simulation prompt card asked what IDEAL assumes about how ethanol and water interact, and whether convergence confirms the property method is appropriate. A student who wrote 'no' to the second question before running arrived at the simulation already skeptical. The convergence without error did not reassure them - it confirmed the question. WHAT MAKES THIS STRONG: Three elements are present. First, the assumption is named specifically - IDEAL, activity coefficients of 1.0, what that physically means. Second, the failure condition is located precisely - near the azeotrope composition, not generically 'when conditions change.' Third, the last sentence resolves the Q4 pre-simulation prompt explicitly: convergence confirmed the calculation completed, not that the method is appropriate. That sentence is doing real intellectual work. THE FAILURE CONDITION TEST: Ask whether the failure condition named could apply to any simulation or only this one. 'If the system behaves differently than expected' fails the test. 'Near the azeotrope at 89 mol% ethanol at atmospheric pressure' passes it. The more specific the failure condition, the more certain you are the student engaged with this system rather than producing a generic answer. NOTE ON AZEOTROPE: Ethanol-water forms a well-known homogeneous azeotrope at ~89 mol% ethanol, 78.2 C at 1 atm. IDEAL cannot predict this - it predicts monotonic behavior. This is verifiable and the student would discover it in HW6 when generating T-xy diagrams with NRTL. |
Section 3.0 - Observed Results¶
| 3.0 - Observed Results SYSTEMS |
| The Rachford-Rice equation governs the flash calculation in FLASH2. At 200 F and 14.7 psia, the adiabatic flash produced approximately 35 mol% vapor. That vapor is enriched in ethanol relative to the feed - ethanol's normal boiling point (173 F) is lower than water's (212 F), so ethanol preferentially partitions to the vapor phase. The stream table confirmed this: the vapor product showed approximately 58 mol% ethanol versus 30 mol% in the feed entering the flash unit. The equation predicted a direction - lighter component goes to vapor - and the stream table confirmed the magnitude. What connected for me: the FLASH2 result is not a black box. It is solving the same phase equilibrium I would calculate by hand with K-values, just iterating to convergence faster than I could. |
| SCAFFOLDING ANNOTATION |
HOW THE STUDENT ARRIVED HERE: The stream table is the source. A student who looks at the vapor product composition and compares it to the feed composition sees the enrichment directly. The connection to Rachford-Rice comes from asking 'what equation is Aspen solving here?' - which is the question Section 3.0 is designed to prompt. WHAT MAKES THIS STRONG: The student names an equation (Rachford-Rice), connects it to a specific numerical output (35 mol% vapor, 58 mol% ethanol in vapor), and traces that number back to a physical property (boiling point difference). The last sentence - 'it is solving the same phase equilibrium I would calculate by hand with K-values' - is the most important line. It demystifies the simulation. A student who understands this will not be confused when FLASH2 fails to converge, because they know what convergence requires. WHAT FIELD 03 IS REALLY ASKING: Not 'name an equation you used' but 'find the place where a mathematical relationship produced a physically observable result.' The numbers should come from the student's own stream table - which means they must have looked at it, not just run the simulation and moved on. MINIMUM STANDARD: Section 3.0 must contain at least one specific number from the simulation output and one connection to a physical property or observable. An entry that names an equation without a number, or a number without a physical connection, is incomplete. |
Section 4.0 - Warrant¶
| 4.0 - Warrant TRADEOFF |
| A skeptical engineer would challenge the SEP block specification - 95% ethanol recovery - and ask for evidence that this split is achievable at these conditions. The SEP block will produce that split regardless of thermodynamics. But at 80 F and 14.7 psia, a 30 mol% ethanol feed is entirely liquid - there is no vapor phase to separate to. A real separation at these conditions would require a different mechanism entirely: distillation, liquid-liquid extraction, or adsorption. The SEP block hid this problem. If I had used a FLASH2 block instead of a SEP block for the first separation, Aspen would have told me that no vapor forms at these conditions - which is physically correct. I used SEP because it was computationally convenient. The skeptical engineer would be right to ask what physical equipment corresponds to that block. |
| SCAFFOLDING ANNOTATION |
HOW THE STUDENT ARRIVED HERE: This requires the student to ask what physical equipment the SEP block represents. SEP is often described in courses as a 'black box separator' - which is accurate but conceals the question of whether the separation it performs is physically realizable. A student who asks 'what piece of equipment does this correspond to?' at 80 F and 14.7 psia discovers that the feed is entirely liquid, and a 95% ethanol split from an entirely liquid stream has no obvious physical mechanism. WHAT MAKES THIS STRONG: The challenge is specific and self-critical. The student is not challenging a classmate's work or a generic simulation practice - they are identifying a problem in their own flowsheet and naming what it would have taken to catch it (using FLASH2 instead of SEP). This is the highest form of Section 4.0: a skeptical challenge that the student can substantiate and that points to a concrete alternative. THE FIELD 04 TEST: The skeptical challenge must name something specific that would actually change the conclusion or require additional evidence. 'A skeptical engineer might question the accuracy of my results' fails the test - it applies to every simulation ever run. 'A skeptical engineer would ask what physical equipment corresponds to the SEP block at these conditions' passes it, because it identifies a specific gap and implies a specific check. NOTE FOR INSTRUCTORS: Some students will write Section 4.0 as a question they already answered - 'a skeptical engineer might question whether IDEAL is appropriate, but I already addressed that in Section 2.0.' This is a sign that the student is treating Section 4.0 as redundant. Prompt them to find a challenge that goes outside the simulation boundary - something the model cannot address, not something the student already addressed. |
The Entry - Weak Version¶
The following entry covers the same simulation. All four fields are present. The simulation was run correctly. The thinking is absent.
| WEAK VERSION - 1.0 - Prediction |
| I learned about the difference between the SEP block and the FLASH2 block. The SEP block separates based on specified fractions while the FLASH2 block uses phase equilibrium. I also learned that the IDEAL property method may not be the best choice for all systems. This was an interesting simulation that helped me understand separators better. |
| WHY THIS FAILS |
This field summarizes course content rather than recording a belief that changed. 'I learned about the difference between SEP and FLASH2' is a content summary - it says the student now knows something they did not know before, but does not name what they believed before or what specific observation changed it. The last sentence ('this was an interesting simulation') is the weakest possible closing - it signals that the student has run out of things to say and is filling space. The test: what did the student believe before this session that they no longer believe? That question cannot be answered from this entry. The field fails the basic definition of a belief update. |
| WEAK VERSION - 2.0 - Simulation Setup |
| I assumed the IDEAL property method was appropriate for this simulation. This assumption could fail if the system behaves differently than expected or if the conditions change significantly. The property method is important to choose correctly for accurate results. |
| WHY THIS FAILS |
The assumption is named (IDEAL) but not examined. The failure condition - 'if the system behaves differently than expected' - applies to every simulation ever run. It contains no specific information about this system, these components, or these conditions. The closing sentence restates the general importance of property method selection without saying anything about why IDEAL specifically is problematic for ethanol-water. The minimum requirement for Section 2.0: name the specific condition under which this assumption fails for this system. For IDEAL and ethanol-water, that condition is near the azeotrope composition. A student who cannot name the azeotrope did not engage with what IDEAL assumes or what ethanol-water actually does. |
| WEAK VERSION - 3.0 - Observed Results |
| The flash calculation uses thermodynamic equations to determine the vapor and liquid compositions at the given conditions. The results showed that ethanol was enriched in the vapor phase because it is more volatile. The equations Aspen uses are based on phase equilibrium principles from thermodynamics. |
| WHY THIS FAILS |
No equation is named. No number from the simulation appears. The observation that ethanol is enriched in the vapor phase is correct but generic - it follows from the boiling point difference without requiring the student to look at the stream table. 'The equations Aspen uses are based on phase equilibrium principles' is true of every flash calculation ever run - it contains no information specific to this simulation. The minimum requirement for Section 3.0: one specific number from the simulation output, connected to one physical property or observable. Without numbers from the stream table, there is no evidence the student looked at the output rather than inferring the result from general knowledge. |
| WEAK VERSION - 4.0 - Warrant |
| A skeptical engineer might question whether the IDEAL property method gives accurate results for this system. They might also question whether the simulation setup is correct. To address this, a better property method could be used in future simulations. |
| WHY THIS FAILS |
The skeptical challenge names something the student already addressed in Section 2.0 - the IDEAL property method - without adding anything new. The second sentence ('they might also question whether the simulation setup is correct') is so broad it is meaningless. The closing sentence proposes a future action rather than identifying a specific gap in the current analysis. A skeptical challenge that can be answered with 'I already addressed that' is not a challenge - it is a restatement. Section 4.0 should go somewhere the simulation cannot follow: the physical realizability of the SEP block, the energy accounting for the separation, the regulatory constraint on what happens to the product stream. Somewhere outside the boundary of what was modeled. |
Evaluation Notes¶
This section is instructor-facing. It is not distributed to students. Use it to calibrate your reading of student entries and to develop the Warrant rubric over time. The strong and weak versions above define the boundaries. The notes below explain what to look for at each criterion.
Criterion 1 - Specificity of Prediction (Section 1.0)¶
The central question: can you identify what the student believed before this session and what specific observation changed it? If either element is missing, the entry fails this criterion regardless of length or sophistication of language.
| Level | What it looks like | Common failure mode |
| Strong | Names prior belief, specific observation that changed it, and consequence of the change | No prior belief named - summary of content only |
| Adequate | Names prior belief and observation but not the consequence | Prior belief is too vague to be falsifiable |
| Weak | Names one element only - usually the new knowledge without the prior belief | Entry is a learning summary: 'I learned that X' |
| Missing | Field is blank, restates the simulation context, or is fewer than two sentences | Field is present but contains no belief content |
Criterion 2 - Precision of Setup Assumptions (Section 2.0)¶
The test: does the failure condition named apply specifically to this system at these conditions, or could it apply to any simulation? Generic failure conditions ('if conditions change,' 'if the system behaves differently') score at the weak level regardless of how well the assumption is named.
| Level | What it looks like | Common failure mode |
| Strong | Assumption named with physical basis; failure condition locates the specific regime or composition where it breaks | Failure condition is generic: 'if assumptions are violated' |
| Adequate | Assumption named; failure condition is a category rather than a specific condition | Assumption is named without physical basis - just the label |
| Weak | Assumption named but failure condition is generic or absent | Failure condition restates the assumption: 'fails if IDEAL is wrong' |
| Missing | Field blank or names a condition rather than an assumption with a failure condition | Student describes what they did rather than what they assumed |
Criterion 3 - Quantitative Grounding of Observed Results (Section 3.0)¶
The minimum standard: at least one number from the student's stream table or block report, connected to a physical property or observable. An entry without numbers has no evidence the student examined their output. An entry with numbers but no physical connection is arithmetic without understanding.
| Level | What it looks like | Common failure mode |
| Strong | Equation named; specific number from output; physical property or mechanism that explains the number; demystifies what Aspen is computing | Number present but no equation and no physical connection |
| Adequate | Number from output connected to a physical observation; equation may or may not be named | Equation named but number comes from general knowledge, not stream table |
| Weak | Number present but not connected to physical meaning, or connection made without a number | Result described qualitatively: 'ethanol was enriched in the vapor' |
| Missing | No numbers; field describes what flash calculations do in general | Student paraphrases lecture content rather than reporting simulation output |
Criterion 4 - Specificity and Reach of Warrant (Section 4.0)¶
Two tests apply. First: is the challenge specific to this simulation, or generic? Second: does the challenge go somewhere the simulation cannot follow - outside the model boundary - or does it only challenge what the student already addressed? The highest-scoring entries identify something the model omits entirely, not just something it approximates.
| Level | What it looks like | Common failure mode |
| Strong | Specific challenge that goes outside the model boundary; names what evidence would be required; student can substantiate the challenge | Challenge restates Section 2.0 without adding new content |
| Adequate | Specific challenge within the model; names what would need to change for the conclusion to hold | Challenge is specific but already addressed elsewhere in the entry |
| Weak | Challenge is generic but not circular - applies to this class of simulations rather than all simulations | Challenge identifies a limitation without naming what it would take to address it |
| Missing | Challenge is completely generic ('a skeptical engineer might question my accuracy') or field is blank | Challenge is a question the student already answered in Fields 01-03 |
Evaluation Iteration Guidance¶
After reading the first cohort of entries, return to these tables and note which level boundary is hardest to call consistently. The most common ambiguity is between Adequate and Weak in Section 1.0 - where the prior belief is named but too vague to be falsifiable. Updating the 'common failure mode' column with specific examples from student entries will improve inter-rater reliability if TAs are grading.
A second common ambiguity is in Section 4.0: entries that name the same limitation as Section 2.0 (property method concern) but frame it as a challenge rather than an assumption. These should score at the Adequate level - the content is not new but the framing shows the student understands that limitations have consequences. Reserve Strong for entries that genuinely go outside the model boundary.
Update this document at the end of each semester with two or three anonymized student entry excerpts at each rubric level. After two or three cohorts, the rubric tables will be calibrated against real student work rather than hypothetical descriptions.
Field Notes for Engineering - engineered-judgment.github.io/field-notes