For classes · EO-1 reference model v1.1
Class assignments on the EO-1 reference model
Six assignments for a systems engineering course, each with a task, what to hand in and a rubric. All of them work on the EO-1 Open MBSE Reference Model, a fictional 6U Earth observation CubeSat, as it ships. Students can answer every question in Holarch or from the model's CSV files in any tool. Free to use and adapt under CC BY 4.0.
On this page
How to use them
The assignments follow the model from stakeholder needs to verification and change. Hand them out in order or pick the ones that fit your course. Each one names the learning outcomes it assesses, a suggested group size and a rubric on four levels with weights; change the weights and levels to match your grading scheme.
- Give every student or group the model. Download EO-1 Reference Model v1.1, or let each group create it in Holarch with Manage Projects → Project Files ▾ → New Demo Project → CubeSat (Guided systems engineering). Every assignment starts from the model as it ships; if a group's project has changed since, have them work in a fresh copy.
- Set up the class. Put the group projects in a class organization, add yourself and your teaching assistants to each group project as Reviewer, and add each assignment's due date in Class Review. The instructor guide has a setup checklist and three module plans.
- Grade the reasoning, not the tool skill. The rubrics reward findings, decisions and their justification. A student who works from the CSV files can reach the top level on every criterion except the ones that ask for changes in the model.
All names, people, organizations and results in EO-1 are invented. EO-1 is not a real spacecraft.
Learning outcomes
The outcomes describe systems engineering ability. They follow the technical processes of ISO/IEC/IEEE 15288 (stakeholder needs and requirements definition, system requirements definition, architecture definition, verification and validation); map them to your program's accreditation criteria as your school requires.
| Outcome | Students can | Main assignment | Also in | Holarch features used |
|---|---|---|---|---|
| LO1 | Tell stakeholder needs from system requirements, and trace each need to the requirements that formalize it. | A1 | A3, A6 | Statements, traced to links, the Item View and its Trace section, the Traceability Matrix |
| LO2 | Write requirements that are necessary, singular, unambiguous and verifiable, with a verification method and success criteria. | A2 | A3 | Requirement attributes, Quality, the TBD/TBR Register |
| LO3 | Run a review and approval cycle on requirements, and answer review comments with changes to the model. | A2 | The Reviewer role, Approvals (Approve, Reject…, Reopen), comments | |
| LO4 | Analyze traceability and verification coverage, and decide for each gap whether the model or the requirement is wrong. | A3 | A1, A4 | Traceability Matrix, RVTM, Model Checks |
| LO5 | Decompose functions, define a logical architecture, allocate functions to logical and physical elements, and derive the interfaces that follow. | A4 | Action diagrams, the Asset hierarchy, allocation, the Interface Register | |
| LO6 | Assess the impact of a requirement change through the model, and record a decision for each affected item. | A5 | Links to Review, impact analysis, Change Requests and their Impact lists, baselines | |
| LO7 | Plan verification and validation: tell them apart, choose methods, write validation criteria in the customer's terms, and read an RVTM. | A6 | A1, A3, A5 | Test cases and suites, Test Center, RVTM and VCRM, document templates |
| LO8 | Produce engineering deliverables from the model and keep them consistent with it: baselined documents and live matrices exported for review. | A6 | A2 | Documents, live matrices, baselines, Word export |
| LO9 (optional) | Read and write SysML v2 textual notation for a subsystem, and state which properties a tool checked and which it only kept. | A6 part B | A SysML v2 project, the Editor's checks, the teaching boundaries below |
A1 · Trace a need
Outcomes: LO1 (main), LO4, LO7 · Group size: individual · Suggested weight: 8 %
EO-1's customer, an agricultural analytics company, stated three needs. N2 Twice-Weekly Revisit asks for fresh imagery of the customers' fields at least twice a week in the growing season. A need is met only when the requirements that formalize it are satisfied by the design and verified, and the need itself is validated with the customer.
Task
- Follow N2 through the model: the requirements it traces to, their children, the functions and components that satisfy each one, and the test cases that verify them, with each test's status.
- Draw the chain as a table or a small diagram: one row or box per item, with its number, name, status and the relationship that connects it to the next.
- Find the validation activity for N2 and its status.
- Answer: what evidence is still missing before N2 counts as validated? List each missing piece, the item it belongs to, and what would close it.
- Look for requirements that support N2 but do not trace to it. Name one candidate and argue for or against adding the trace. Do not add it to the model.
In Holarch: open N2's Item View and follow its Trace section, then the Traceability Matrix, the RVTM and Verification → Test Center. Without Holarch: eo-1-v1.1-traceability.csv and eo-1-v1.1-rvtm.csv.
Hand in
A note of one to two pages (Word or PDF) with the chain, the missing evidence and your answer to step 5. Do not change the model for this assignment.
Rubric
| Criterion | 4 Exemplary | 3 Proficient | 2 Developing | 1 Beginning |
|---|---|---|---|---|
| Chain complete and correct (40 %) | Every item from N2 to the tests, with the right relationships and statuses, including every child requirement and the validation activity. | One item or status missing or wrong. | The chain stops at the requirements or misses a child. | Items listed without relationships. |
| Missing evidence (35 %) | Names every gap (tests not run, requirements not yet approved, a requirement without a satisfying function, validation not run) and what closes each. | Most gaps, with what closes them. | Lists gaps without saying what closes them. | Says only that tests are missing. |
| Verification and validation (15 %) | Keeps the two apart throughout: requirements are verified, the need is validated. | One slip. | Treats a passed test as validating the need. | No distinction. |
| Candidate trace (10 %) | A specific requirement with a reasoned argument either way. | A candidate with a thin argument. | A candidate without an argument. | Missing. |
A2 · Requirements rework and review
Outcomes: LO2 (main), LO3, LO8 · Group size: group · Suggested weight: 10 %
Three EO-1 requirements are still drafts and score 44 % in Holarch's quality check:
- 2.5 Safe Mode Recovery: "The spacecraft should recover from safe mode quickly and resume imaging as appropriate"
- 3.5 Battery Thermal Control: "The thermal design shall keep the battery warm enough (TBD)."
- 6.4 Operator Interface: "The operator interface shall be user friendly and/or easy to use."
Task
- For each of the three, list what makes it weak, word by word where it matters.
- Rewrite each in your group project: one "shall", one measurable condition, and a stated starting point or condition where time or range is involved. Replace every TBD with a value and record its rationale, or keep a TBR with an owner and a due date.
- Fill in Verification Method, Verification Level and Success Criteria for each.
- Trace each to where it comes from (a need, a parent requirement or a source) and to the function that satisfies it. If no function exists, say which function is missing.
- Set each to In Review. Your instructor reviews them as a Reviewer, with comments and Approve or Reject…. Answer every comment in the model: change the requirement, or add a comment that gives your reason. Rework a rejected requirement and set it to In Review again.
- Save a baseline of the System Requirements Specification named "A2 handed in" and export it to Word.
In Holarch: Needs & Requirements → Quality, the requirement's inspector, Needs & Requirements → Approvals, the document's Version ▾ → Save Baseline… and Deliver ▾ → Specification (Word). Without Holarch: eo-1-v1.1-requirements.csv; hand in the rewrites as a table, and your instructor plays the review round by email or in class.
Hand in
The Word export, and half a page on the weaknesses you found in step 1. The model as it stands at the due date is graded with it.
Rubric
| Criterion | 4 Exemplary | 3 Proficient | 2 Developing | 1 Beginning |
|---|---|---|---|---|
| Rewritten requirements (40 %) | All three singular, measurable and verifiable; conditions and starting points stated; values justified. | One requirement still has a vague term or a missing condition. | Values added, but conditions or rationale missing. | Wording changed, still not verifiable. |
| Verification attributes (20 %) | Method, level and success criteria fit each requirement and could be carried out as written. | One mismatch, for example Test where only Analysis is possible. | Methods set, success criteria vague. | Missing. |
| Traceability (15 %) | Each traced to its source and its satisfying function, or the missing function named. | One trace missing. | Sources only. | None. |
| Review cycle (15 %) | Every review comment answered in the model; rejected requirements reworked and resubmitted. | One comment left unanswered. | Comments answered without changes to the model. | Comments ignored. |
| Weakness analysis (10 %) | Precise: names each vague term and why it cannot be verified. | General but correct. | Restates the quality score. | Missing. |
A3 · Find the coverage gaps
Outcomes: LO4 (main), LO1, LO2, LO7 · Group size: group · Suggested weight: 10 %
EO-1 is at its critical design review. The model is deliberately imperfect: some requirements have no function that satisfies them, some have no test, some tests failed, and one equation is linked to nothing.
Task
Work from the model as it ships. Use a fresh copy if your group project has changed.
- List every requirement that no function satisfies.
- List every requirement that no test case verifies, and those whose tests failed, with the open issue for each failure.
- List the requirements with no satisfying component. Decide which of them need one and which are rightly satisfied at mission or segment level.
- Find model items that nothing traces to. Hint: look at the power equations.
- For each gap, decide: is the model missing something (a function, a component, a test, a link), or is the requirement not needed or wrongly written? Give one sentence of reasoning per gap.
- Close three gaps of your choice in your group project, each as one change you can point to in Control & Reviews → Activity.
In Holarch: the Traceability Matrix, the RVTM and Verification → Model Checks. Without Holarch: eo-1-v1.1-traceability.csv (its Coverage column) and eo-1-v1.1-rvtm.csv; describe the three changes instead of making them.
Hand in
A gap table (gap, item, decision, reasoning, closed yes or no) as Excel or Word, and a list of the three changes you made.
Rubric
| Criterion | 4 Exemplary | 3 Proficient | 2 Developing | 1 Beginning |
|---|---|---|---|---|
| Gaps found (35 %) | All function, test, component and untraced-item gaps, including the failed tests and their issues. | One or two gaps missed. | Only the obvious drafts found. | Fewer than half. |
| Decisions (35 %) | Each gap has a defensible decision; missing design is told apart from unneeded or mis-scoped requirements and from mission-level satisfaction. | Most decisions sound. | Every gap treated as "add a link". | No decisions. |
| Changes made (20 %) | Three changes that really close gaps, each one clean step in the record. | Two good changes. | Changes that hide a gap, such as a link added without a reason. | None. |
| Presentation (10 %) | The table is sortable and complete. | Minor omissions. | Hard to follow. | Missing. |
A4 · Extend the architecture
Outcomes: LO5 (main), LO4 · Group size: group · Suggested weight: 12 %
EO-1's logical architecture, L0 EO-1 Logical Architecture, covers only the power and data handling slices: L1 Power Generation to L6 Fault Management. Attitude determination and control has the functions F4.1 Estimate Attitude, F4.2 Compute Control Torques, F4.3 Drive Reaction Wheels and F4.4 Unload Wheel Momentum, and physical parts under 1.2 ADCS, but no logical layer.
Task
- Add logical elements for attitude determination and control under L0. Each element states its purpose without naming hardware.
- Allocate F4.1 to F4.4 to them, and relate each logical element to the physical parts under 1.2 ADCS that realize it. Follow the pattern of L1 to L6: open one of them in the Item View and look at its Trace section to see which relationships it uses.
- Draw the action diagram of F4 with the items that flow between its subfunctions, for example the attitude estimate and the torque commands, if your group project does not have one yet.
- From the flows that cross element boundaries, derive the interfaces. Compare them with the existing interfaces IF-06, IF-16, IF-17 and IF-18: which match, which are missing, and which carry items no function uses?
- Check that 2.1 to 2.4 are each satisfied by the function that does the work, and say which logical element is then responsible for each requirement.
In Holarch: the Asset hierarchy, the action diagram of F4, the Interface Register, and a Hierarchy diagram with the Allocation Hierarchy preset. Without Holarch: eo-1-v1.1-allocation-chain.csv, eo-1-v1.1-icd.csv and eo-1-v1.1-traceability.csv; draw the diagram by hand or in any tool.
Hand in
A short architecture description (two to three pages, Word or PDF): the logical elements with their purposes, the allocation table (function, logical element, physical part), the F4 action diagram and the interface comparison.
Rubric
| Criterion | 4 Exemplary | 3 Proficient | 2 Developing | 1 Beginning |
|---|---|---|---|---|
| Logical elements (25 %) | Solution-independent and cohesive, each with a clear purpose; no hardware names. | One element names hardware or mixes purposes. | Elements copy the physical parts. | Missing or unclear. |
| Allocation (30 %) | Every subfunction allocated once to the right element; elements realized by parts; consistent in the model. | One allocation questionable. | Allocations only to parts, skipping the logical layer. | Incomplete. |
| Flows and interfaces (30 %) | Interfaces derived from the flows that cross boundaries; a precise comparison with IF-06, IF-16, IF-17 and IF-18, including gaps both ways. | Comparison mostly right. | Interfaces listed, not derived. | Missing. |
| Requirement responsibility (15 %) | Each of 2.1 to 2.4 traced through its function to a logical element, with inconsistencies named. | One missing. | Listed without checking. | Missing. |
A5 · Work the change
Outcomes: LO6 (main), LO7 · Group size: group · Suggested weight: 10 %
After the design review, requirement 3.1 Orbit Average Power changed from 28 W to 30 W orbit-average power at end of life. Test case TC-06 Solar Array Power Test passed against the old value. The links from 3.1 are now listed under Links to Review.
Task
- List every item the change reaches: forward through the links from 3.1, backward to what 3.1 derives from, and items that mention the old value without a link to 3.1. Say how you found each one.
- For each item, write the question the change raises, your finding (with numbers where the model has them) and a decision: no change (mark the link reviewed, with a comment that gives the reason), update, or escalate to the change board.
- Decide what to do with TC-06: does a test that passed at 28 W still prove 30 W? Show the arithmetic from the model's equations.
- Find one item that the link-based analysis misses, and say why it matters.
- Record your decisions in your group project. Save a project baseline before you start. Mark the links that still hold as reviewed, with comments. Create a change request for what the board must decide (the model already holds CR-1 as a draft, so yours is CR-2), add the affected items, and stage the changes you propose. Read each staged change's Impact list and compare it with your own record: what does it list that you missed, and what did you find that it cannot list? Submit the request for review without applying it, then save a second baseline.
Optional: if your instructor turns on Work in change requests in the project's change request Settings, make your proposed edits in the draft request on the usual pages instead of staging them one by one, and review them in the request before you submit it.
In Holarch: Needs & Requirements → Links to Review, an impact analysis from 3.1 (Open in ▾ → Spider, then Impact), the Item View's Trace section, search, the power equations and TPM-2, Control & Reviews → Change Requests and Control & Reviews → Baselines & Versions. Without Holarch: eo-1-v1.1-traceability.csv for the analysis; eo-1-v1.1-change-impact-3.1.csv is the model's own impact record, so do your analysis first and compare afterwards.
Plans: change requests, their Impact lists and approval records work on every plan. Configuration control (approved and baselined items change only through an approved request) comes with the Individual and Team plans, and working in a change request with the Team plan. A verified academic email address or membership in an active class organization includes both. See Configuration control.
Hand in
An impact record (a table: item, how it was reached, question, finding, decision, status) and a one-paragraph recommendation to the change board.
Rubric
| Criterion | 4 Exemplary | 3 Proficient | 2 Developing | 1 Beginning |
|---|---|---|---|---|
| Reach (30 %) | Forward, backward and text-search items found, including those no link reaches. | Links only, all of them. | Some forward links. | 3.1 and TC-06 only. |
| Findings with numbers (30 %) | Uses the model's values and equations; the effect on TC-06 is computed and explained. | Numbers mostly right. | Qualitative only. | Wrong conclusions. |
| Decisions (25 %) | Each item has a decision that fits the finding; links marked reviewed only with a reason; board items kept separate. | One decision weak. | Links marked reviewed without reasons. | No decisions. |
| Record in the model (15 %) | Comments, the change request with its staged changes and the two baselines in place and consistent with the record; differences from the Impact list explained. | One missing. | Only the written record. | Nothing in the model. |
A6 · Validation plan
Outcomes: LO7 (main), LO8, LO1; part B: LO9 · Group size: group · Suggested weight: 10 %
Verification asks whether EO-1 was built right against its requirements; validation asks whether it is the right system for the farm customers. EO-1's validation suite holds VAL-01 Customer Alert Pilot (for N1), VAL-02 Growing Season Revisit Simulation (for N2) and VAL-03 Agronomist Map Review (for N3, passed).
Start the plan with + New Document on the Documents page. Under Templates for, choose your instructor's class template if there is one; otherwise choose the built-in Verification and Validation Plan (a Test Plan Document) with Include writing guidance. The section numbers below are the built-in template's.
Task (part A)
- 5.1 Operational Scenarios: write the scenario for VAL-01 in the customer's terms: who does what, on which day, with which result. Name the need it exercises.
- 5.2 Users and Participants: name the representative users and what each of them decides.
- 5.3 Validation Environment: describe the environment, how it differs from operations, and what that difference means for the result.
- Write VAL-01's acceptance criteria so that a customer could say yes or no without engineering judgment, and record them on the test case.
- 8 Verification Status: read the live RVTM. Pick the failed test TC-04 Pointing Accuracy Test and say what VAL-01's result would mean if TC-04 is still failed at the time of the pilot.
- Replace the writing guidance with your content and export the plan to Word.
Part B: SysML v2 (optional)
Use it as a bonus of up to 10 %, or on its own in a SysML v2 module. Create a SysML v2 project from eo-1-v1.1-power-subsystem.sysml with Manage Projects → Project Files ▾ → Import SysML v2 Files…. The file still holds 3.1 at 28 W.
- Change the required orbit-average power to 30 W, the change from A5, and read what the Editor reports. Is it a modeling error or a design finding?
- Change the design values, not the requirement, until the file has 0 diagnostics. Say whether your new values are physically credible.
- In half a page, state which of your claims the tool checked and which it only kept. Use the teaching boundaries below.
Without Holarch: write the plan in Word from the section headings above; take the verification status from eo-1-v1.1-rvtm.csv.
Hand in
The Word export of the plan, and for part B the .sysml file and the half page.
Rubric
| Criterion | 4 Exemplary | 3 Proficient | 2 Developing | 1 Beginning |
|---|---|---|---|---|
| Scenario in the customer's terms (30 %) | Concrete and day-in-the-life, without engineering jargon; clearly exercises N1. | Clear, with some jargon. | A test procedure, not a scenario. | Missing or generic. |
| Acceptance criteria (25 %) | Observable, yes or no, owned by the customer, and tied to what N1 asks for. | Mostly observable. | Engineering thresholds only. | Vague. |
| Participants and environment (20 %) | Representative users named with their decisions; differences from operations and their effect stated. | One missing. | Listed without their effect. | Missing. |
| Verification and validation reasoning (15 %) | Explains why a validation pilot can succeed or fail independently of TC-04, and what each outcome means. | Mostly sound. | Confuses verification and validation. | Missing. |
| Deliverable (10 %) | A clean Word export with the writing guidance replaced by content. | Leftover guidance text. | Partly exported. | Missing. |
| Part B: SysML v2 (bonus, up to 10 %) | The finding explained; 0 diagnostics with credible values; a precise statement of what was checked and what was kept. | 0 diagnostics; the statement is general. | The requirement weakened instead of the design changed, or diagnostics left. | Not attempted. |
Working without Holarch
Every assignment can be done from the EO-1 files alone, so students who use another tool, or none, can take part. The CSV files are UTF-8 with a header row and open in any spreadsheet. They are views of the same model, version 1.1.
| Assignment | Files |
|---|---|
| A1 Trace a need | traceability, RVTM |
| A2 Requirements rework | requirements, traceability |
| A3 Coverage gaps | traceability (Coverage column), RVTM, issues and actions |
| A4 Architecture | allocation chain, interfaces, traceability |
| A5 Work the change | traceability; afterwards, to compare: change impact for 3.1 |
| A6 Validation plan | RVTM, requirements; part B: power subsystem (.sysml) |
All files, with a README and the license, are in eo-1-v1.1.zip. The EO-1 downloads describe each file.
Teaching SysML v2: what Holarch checks and what it only keeps
Holarch's SysML v2 projects support a subset of the OMG SysML v2 specification. When you teach the language with it, keep three kinds of statement apart, so students neither learn a tool limit as a language rule nor take a construct the tool kept as one it checked. The SysML v2 language coverage page lists the current state; check it before each term.
Checked
The Editor checks these as you type, and the views and pages use them. When the Editor shows 0 diagnostics, these checks passed.
- Syntax: characters, comments, strings, braces,
;, expected names and tokens. - Names and imports: unresolved, ambiguous, duplicate and private names; membership, namespace, recursive and filtered imports.
- Definitions, usages and specialization: typing, subsetting, redefinition and conjugation; cycles and the wrong kind of definition.
- Multiplicities, values and units: valid bounds, quantities with the standard library's ISQ, SI and US customary units, and dimension checks (a power where a mass is expected is reported).
- Constraints and calculations: operand types, calculation arguments and Boolean conditions, and evaluation: constraints, requirements and objectives that do not hold for the current values are reported.
- Structure: connection ends and the directions of ports, flows and messages.
- Requirements, use cases and views: members in the right place, and
satisfy,verify,exposeandframetargets. - Actions and states: successions, control nodes, initial transitions, parallel regions, triggers and effects. The Simulator runs action flows, including Monte Carlo runs, and runs state machines with stimuli.
- Variability, individuals and time slices: a variant has its variation's features and a choice must be one of its point's variants; a snapshot or time slice has the features of its occurrence.
Checked, with a limit to name
| Area | Limit to tell students |
|---|---|
| Analysis cases | An analysis case is evaluated only when every input has a value; there is no solver. In a Parametric View, Solve finds one unknown per constraint. Trade-study analyses from the standard library are not evaluated. |
| Variability | Variations, variants and configurations are checked and evaluated, but Holarch's pages do not list every valid combination yet. A variation point without a choice has an unknown value, so an assertion over it is neither met nor violated. |
| Metadata and language extensions | User-defined keywords work through semantic metadata. Other user-defined metadata has no meaning beyond its values. The standard library is read-only. |
| KerML | The standard library's KerML is read, but .kerml files are not imported and models written in KerML are not checked. |
| SysML v2 JSON exchange | Element kinds Holarch does not interpret are kept as imported and written back unchanged on export; they have no text and cannot be edited. |
| Systems Modeling API | Other tools read the model and commit into change request branches; an approver applies them in Holarch. There are no commits to main, tags or merges. |
Kept, not interpreted
Holarch keeps these exactly as written, in the text and in exports. 0 diagnostics says nothing about whether they are right.
- Textual representations in other languages (
rep language "Python"and the like): the body is kept as text and never run. Bodies in SysML and KerML are checked. - The words of a requirement: the
doctext is the statement a reader sees; the check evaluates therequire constraint, not the words. In the EO-1 power subsystem, changing the required value to 30 W leaves thedoctext saying 28 W until someone edits it. It makes a good exercise. - The values of user-defined metadata other than semantic metadata.
How to use this in a course
- Say "the tool checked it" only for the checked areas. For example, 0 diagnostics means the units are consistent and the requirement holds for the current values. It does not mean the requirement text agrees with its constraint, or that a body written in another language is correct.
- Assess the kept constructs by reading them, not by the diagnostics count.
- Language rules come from the specification, not from any tool. Where Holarch has not implemented something, say "not yet in Holarch", not "not allowed".
- Part B of A6 builds on this section: students see the evaluation report the changed requirement, fix the design and write down what was checked and what was only kept.
Solutions
Instructors can request the solution notes (expected findings, sample answers and common mistakes for A1 to A6) through support: contact us with the topic Instructor, your course and your school. The notes are not published, so students cannot download the answers.
License and credit
These assignments are licensed under the Creative Commons Attribution 4.0 International License (CC BY 4.0). You may copy, hand out and adapt them, if you give appropriate credit, link to the license and say whether you made changes. Credit: "Class assignments on the EO-1 reference model, Holarch, CC BY 4.0". The EO-1 model files keep their own credit line.
The license covers the assignment texts only. The Holarch software is not covered by it.