Simulations for Engineering Education: How to Assess Technical Problem-Solving, Design Justification, and ABET Competencies with AI

Engineering education prepares graduates to design systems that protect human life, infrastructure, and economic stability. Before an engineer approves a foundation, certifies a pressure vessel, or deploys a control system, faculty must verify that students can analyze physical constraints, calculate safety margins, and defend design trade-offs under scrutiny.

For generations, engineering departments evaluated competence through static deliverables: problem sets, calculation sheets, and design reports. A professor assigned a design specification. Students gathered material properties, applied governing equations, and submitted technical dossiers.

Generative artificial intelligence disrupted that evaluation model. Multimodal foundation models now ingest design prompts and generate engineering reports in seconds, outlining equations, solving calculations, writing scripts, and formatting summaries.

When software generates reports on demand, static documents stop proving individual mastery. Competence does not exist in generated text. Professional practice requires physical reasoning, technical intuition, and defending decisions under unexpected conditions. Engineering programs need assessments that evaluate technical reasoning in action, which explains why simulations for engineering education and authentic assessment are transforming higher education.

Why Written Engineering Reports Fall Short

The vulnerability of written calculation packages extends beyond software concerns. Traditional deliverables exhibit structural weaknesses that generative tools exposed.

First, static reports reward plausible text over physical intuition. Large language models synthesize authoritative prose that conceals flawed assumptions. A student can submit a generated calculation package containing impossible boundary conditions, such as negative thermal resistances or unconstrained stresses. Identifying whether the student understands physical laws requires forensic inspection.

Second, written deliverables isolate students from design friction. In textbook problems, variables appear clean. In practice, operating data is noisy, specifications conflict, and field parameters drift. Practicing engineers must discover missing parameters through diagnostic inquiry while balancing competing physical constraints.

Third, accreditation standards demand direct verification of individual performance. The Engineering Accreditation Commission of ABET defines clear expectations under Criterion 3 Student Outcomes. Programs must provide verifiable evidence that graduates can solve complex problems (Outcome 1), apply engineering design under realistic constraints (Outcome 2), communicate effectively with technical audiences (Outcome 3), and recognize ethical responsibilities (Outcome 4). Static papers generated by commercial software cannot demonstrate individual attainment of these competencies.

The Architecture of an Educational Engineering Simulation

Designing an effective engineering simulation requires more than prompting a generic chatbot. Unconstrained models agree with flawed calculations, accept unsafe safety factors, and lack technical boundaries.

A rigorous educational simulation relies on four core architectural components:

1. Guarded Technical Constraints

Authentic engineering involves asymmetric information. Operators know maintenance histories, technicians know vibration anomalies, and clients know budget limits. In Rocketproof, the AI persona holds private operational boundaries. Students must ask diagnostic questions to uncover operational constraints before finalizing calculations.

2. Behavioral Pushback and Resistance Curves

Industry stakeholders question aggressive assumptions. Faculty configure resistance curves governing persona responses. If a student proposes an undersized pump or a marginal safety factor, the persona raises operational objections. When the student cites governing standards and demonstrates safe margins, the persona validates the recommendation, reinforcing skills taught in role-play for communication skills.

3. Physical Failure Modes

Simulations incorporate realistic failure modes, including cavitation, thermal runaway, or fatigue fracture. Students must troubleshoot anomalous sensor readings, determine root causes, and adjust operating parameters to maintain system stability.

4. The Post-Simulation Technical Defense

Following the simulation, the student completes an asynchronous oral defense. The platform asks targeted, source-grounded questions about submitted calculations and decisions. Students verbally explain mathematical choices, defend safety factor selections, and justify how their design satisfies ethical codes, using source-grounded questioning.

Two Concrete Disciplinary Blueprints

To see how engineering departments implement simulations, consider two curricular blueprints deployed in mechanical and civil engineering courses.

Blueprint 1: Mechanical Engineering (Thermal and Fluid Systems)

In an upper-division thermal systems course (MECH 340), students complete an industrial heat exchanger optimization scenario.

• The Persona: Marcus Vance, Senior Reliability Engineer.
• The Situation: River temperatures have risen by eight degrees Celsius, reducing cooling effectiveness on an exothermic reactor cooler.
• The Hidden Information: The pump operates near Net Positive Suction Head limits. Increasing flow by twelve percent causes severe impeller cavitation.
• The Student Task: Calculate heat transfer area using LMTD and NTU methods. Interview Marcus to determine pump curves and fouling factors, then balance thermal duty against pump life.
• The Oral Defense Questions:
  1. "How does increasing coolant velocity impact tube bundle pressure drop, and why does your design avoid pump cavitation?"
  2. "On Page 4 of your calculations, you assumed a constant fouling factor. If fouling doubles over thirty days, when does effluent temperature violate safety limits?"
  3. "Explain the thermodynamic trade-off between installing an auxiliary plate exchanger versus derating reactor production during peak summer ambient temperatures."

Blueprint 2: Civil Engineering (Geotechnical and Structural Design)

In a senior civil engineering design course (CE 420), students evaluate an unexpected subsurface condition during foundation construction.

• The Persona: Elena Rostova, Municipal Geotechnical Inspector.
• The Situation: Excavations for a five-story hospital addition encounter an unmapped karstic void four meters below anticipated bedrock, halting work.
• The Hidden Information: The developer faces delay penalties and pressured the contractor to backfill the cavity with lean concrete without geotechnical verification.
• The Student Task: Review boring logs, recalculate bearing capacities, interview Elena regarding bedrock socket requirements, and propose a remediation plan.
• The Oral Defense Questions:
  1. "Why did you select micropiles socketed below the void over pressure-grouting the cavity, and how did you verify load transfer to competent strata?"
  2. "Walk me through how your proposed foundation redesign meets structural settlement limits defined in ASCE 7."
  3. "The contractor requested certification based solely on probe holes. Cite the ASCE Code of Ethics to explain your professional obligation to public safety and client management."

Designing an Objective Engineering Evaluation Rubric

Grading engineering simulations requires criterion-referenced rubrics. Evaluations must measure technical rigor, physical intuition, and professional responsibility across four objective dimensions:

| Rubric Dimension | Developing (1-2 Points) | Proficient (3 Points) | Advanced (4 Points) | | :--- | :--- | :--- | :--- | | Technical Rigor and Calculations | Misapplies formulas; numerical errors; overlooks boundary effects. | Valid governing laws; accurate calculations; checks boundary conditions and units. | Derives governing relationships from first principles; conducts parameter sensitivity analysis. | | Diagnostic Inquiry and Discovery | Accepts initial assumptions; misses hidden operating constraints. | Asks targeted questions; uncovers hidden boundaries; gathers necessary data. | Probes system boundaries systematically; identifies unstated risks and upstream constraints. | | Engineering Judgment and Trade-Offs | Recommends impractical solutions; struggles to balance competing goals. | Balances safety margins, economic feasibility, and technical performance. | Defends optimal compromises using multi-criteria analysis; accounts for lifecycle failure risks. | | Professional Ethics and Standards | Dismisses safety standards; yields to schedule pressure. | Applies professional codes (ASME, ASCE, IEEE); protects public safety. | Articulates ethical principles under stakeholder pressure; establishes clear documentation trails. |

This structured rubric ensures that evaluations measure verifiable outcomes. Faculty maintain complete control over evaluation criteria, adjusting weights and score thresholds to match course learning goals, reflecting the core principles of faculty-controlled AI assessment.

Scaling Technical Evaluation Without Exhausting Faculty

Department chairs frequently ask how oral simulations can work in large engineering courses. In a thermodynamics course with one hundred and twenty students, conducting individual technical interviews is impossible within standard teaching loads, consuming thirty hours of faculty time.

Rocketproof solves this operational challenge through an asynchronous architecture. Students complete simulations and spoken defense responses independently through their browser or mobile device, providing consistent questioning and realistic persona interactions.

Following submission, the evaluation engine performs administrative review:

1. Transcribes spoken responses with specialized vocabulary recognition for engineering terms, units, and formulas.
2. Extracts evidence from dialogue and submitted calculations, mapping timestamped quotes to each rubric dimension.
3. Flags logical contradictions, unaddressed boundary conditions, or safety violations.
4. Generates an Evidence Summary card presenting highlighted clips, transcripts, and provisional rubric scores.

When an instructor opens the dashboard, they review the Evidence Summary, play twenty-second audio clips where the student defends critical assumptions, confirm or adjust scores, and add feedback in three to four minutes per student.

This workflow integrates with institutional learning management systems through LTI 1.3 standards, syncing assignment prompts, rosters, and gradebook columns with Canvas, Blackboard, Brightspace, and Moodle.

To ensure equitable access, the platform includes Universal Design for Learning features. Students with documented speech accommodations or anxiety disorders can complete the defense using structured text responses, adjustable response timers, and low-bandwidth audio options, following proven principles for accessibility in oral assessment.

How Engineering Faculty Begin

Transitioning to simulation-based assessment does not require rebuilding an entire curriculum. Successful engineering programs follow a five-step implementation roadmap:

1. Select a Capstone or Gatekeeper Course: Begin with a course where students transition from theoretical equations to applied design, such as machine design, fluid mechanics, or senior capstone.
2. Convert an Existing Design Problem: Take an existing assignment and identify the real-world reviewer, such as a plant supervisor, regulatory inspector, or client, following methods detailed in how to turn a case study into an interactive AI simulation.
3. Establish Hidden Constraints: Define two or three realistic parameters discovered through dialogue, such as temperature fluctuations, supply delays, or pressure head limitations.
4. Calibrate Defense Questions Around First Principles: Write question stems targeting why students made specific assumptions, how calculations hold up against physical edge cases, and which ethical standards govern choices.
5. Run a Formative Practice Round: Allow students to experience a non-graded simulation attempt early in the semester. Rehearsal familiarizes students with conversational assessment, reduces performance anxiety, and reinforces technical communication habits.

Frequently Asked Questions

How do engineering simulations prevent students from consulting generative AI during the defense?

Rocketproof separates calculation submission from the spoken defense. The oral defense presents dynamic, source-grounded questions with real-time response windows. Students verbally explain equations, assumptions, and data in their submissions without pause. Reading generated text produces identifiable pauses, disconnections from submission specifics, and superficial explanations that faculty easily detect.

Does building an engineering simulation require computer programming skills?

Faculty do not write code. Instructors create scenarios using a visual interface, describing the persona, uploading rubrics, and setting hidden parameters in plain English. The platform automatically configures the simulation and questioning engine.

How does the system evaluate open-ended engineering designs with multiple valid answers?

Engineering design rarely produces a single correct answer. Sound engineering balances competing factors including cost, weight, and safety. Rocketproof evaluates reasoning against governing physical principles and rubric criteria, allowing diverse valid technical paths. If a student chooses an unconventional design, the defense asks them to justify parameters, prove safety factor compliance, and defend trade-offs.

Can simulations support ABET continuous improvement documentation?

Yes. Rocketproof provides assessment analytics mapped directly to ABET Criterion 3 Student Outcomes. Faculty can export cohort performance data, rubric distributions, and timestamped defense evidence to support ABET self-study reports and continuous improvement reviews, demonstrating verifiable student competencies.

Restoring Professional Judgment to Engineering Assessment

Engineering is an applied discipline where technical knowledge meets physical reality. When artificial intelligence solves textbook equations and formats design memos, evaluating static deliverables no longer verifies whether students possess the judgment required for professional practice.

By combining realistic AI simulations with source-grounded oral defenses, engineering educators evaluate what students actually understand. Students learn to discover incomplete information, defend design choices before skeptical stakeholders, uphold ethical codes, and explain complex technical principles with clarity.

To see how engineering departments implement authentic technical simulations and oral defenses, explore the interactive demo or review scalable authentic assessment strategies for your institution.