Triz Engineering Solver

TRIZ Engineering Solver

A systematic analytical engine for solving engineering problems with Genrich Altshuller's Theory of Inventive Problem Solving (TRIZ). Replaces trial-and-error brainstorming with algorithmic problem-solving over a corpus of patent-derived patterns.

When to use

• Engineering trade-offs where improving parameter A degrades parameter B (technical contradiction)
• A single element must exhibit opposite properties (physical contradiction)
• Design bottleneck where conventional optimization has plateaued
• System redesign aimed at the Ideal Final Result

When NOT to use

• Pure software architecture without a physical analogue
• UX / interaction design
• Business or organizational strategy
• Open brainstorming with no concrete contradiction identified

Inputs required

Before analysis, collect from the user:

1. System description — current engineering system and its primary function
2. Problem statement — the specific bottleneck or undesired effect
3. Contradiction parameters — which parameter must improve, which one degrades
4. Constraints — environmental, cost, manufacturing, physical limits
5. Available resources — substances, fields, geometry, byproducts, environment

If any input is missing, use the self-prompting block in ## Intake prompt below.

Workflow

Step 1 — Formulate the Ideal Final Result (IFR)

Describe the outcome where:

• The desired function is achieved perfectly
• No additional cost, complexity, or harmful side-effect is introduced
• Only existing internal / external / temporal resources are used
• The problem "solves itself"

Question to drive Step 1: "In an ideal world, how would this system achieve [desired function] using only what already exists, with zero added complexity?"

Step 2 — Classify the contradiction

Technical contradiction: improving parameter A causes parameter B to worsen.

• Example: increasing strength (A) increases weight (B).

Physical contradiction: a single element must exhibit opposite properties simultaneously or under different conditions.

• Example: a surface must be rough (for grip) and smooth (for low friction).

Physical contradictions are typically resolved by separation in space, time, condition, or system level (parts vs. whole). For the decision procedure and the principles linked to each separation axis, see resources/separation_principles.md. Technical contradictions are resolved via the Contradiction Matrix → Inventive Principles.

Step 3 — Map to the Contradiction Matrix

1. Identify the improving parameter from the 39 — see resources/39_parameters.md.
2. Identify the worsening parameter from the 39.
3. Look up the cell in resources/contradiction_matrix.json. The cell lists 3–4 recommended inventive principle IDs.
4. For each suggested principle, retrieve its full description and sub-principles from resources/40_principles.md.
5. Apply the principles to the concrete system to generate candidate solutions.

> Note: resources/contradiction_matrix.json contains the full Altshuller 39×39 matrix (1190 populated cells; 292 cells are legitimately empty in the original — Altshuller's patent analysis found no dominant principle for those contradictions). For empty cells, fall back to reasoning over the 40 principles directly and mark the suggestion as Source: inferred rather than Source: matrix. See the JSON meta.primary_source for transcription provenance.

Step 4 — Su-Field (Substance–Field) analysis

Model the system's function as a triad:

• S1 — object being acted upon
• S2 — tool / agent acting on S1
• F — field carrying the action (mechanical, thermal, chemical, electromagnetic, gravitational, acoustic)

Diagnose:

• Incomplete system (missing S2 or F): add the missing element, preferring existing resources.
• Insufficient/ineffective interaction: introduce a more controllable field, intermediate substance, or field additive.
• Harmful interaction: insert a barrier substance, introduce a counteracting field, or modify field properties.

For systematic Su-Field transformations, consult resources/76_standard_solutions.md — class taxonomy, diagnostic flow, and the most-cited operational sub-rules of the 76 Standard Solutions (Altshuller, grouped into 5 classes). When you cite a Standard Solution, mark Source: standard-solution- in the output.

Step 5 — Resource utilization

Enumerate before generating concepts:

• Internal resources: unused properties of components, byproducts, idle time, geometric features (holes, surfaces, edges).
• External resources: environmental substances (air, water, gravity), environmental fields (magnetic, thermal gradients), supersystem (adjacent systems).
• Temporal resources: pre-process preparation, post-process utilization, parallel operations.

Prefer concepts that consume free resources over concepts requiring new components.

Step 6 (optional) — ARIZ deepening

For hard problems where the 40-principles pass yielded no strong concept (no concept reaches ideality > 1, see Output format below), switch to ARIZ-85C (Algorithm for Inventive Problem Solving) — a formal 9-part procedure that reframes the problem mini-problem → operational zone → operational time → substance-field resources → IFR-1 → physical contradiction → standard solution. Full procedure and operational checklist: resources/ariz_85c.md. Invoke explicitly when the user asks for "deep TRIZ" / "ARIZ analysis" or when the quickstart fails the ideality bar.

Step 7 (optional) — Trends of Engineering System Evolution

For prognostic / roadmapping questions ("where will this technology go next?"), map the system to the 8 trends (increasing ideality, non-uniform development of parts, transition to supersystem, transition to micro-level, increasing dynamism, complexity → convolution, matching/mismatching, reducing human involvement). Full list, S-curve framing, and the diagnostic procedure: resources/evolution_trends.md. Out of scope for the contradiction-resolution workflow above — invoke only when the user's question is prognostic, not corrective.

Output format

The full machine-readable template lives in resources/output_template.md — use it verbatim. The sections below are the inline summary for quick reference. Every concept must report a numeric (or banded) ideality estimate (see "Ideality metric" below). Concepts with ideality ≤ 1 are dropped, not compromised.

Generate a structured analysis containing the sections below. Keep each section tight; this is an engineering deliverable, not an essay.

1. Contradiction statement

Technical contradiction: improving [Parameter A, with #] causes [Parameter B, with #] to worsen.
— OR —
Physical contradiction: [Element X] must be [Property 1] AND [Property 2].
  Separation strategy candidate: [space | time | condition | system-level].

2. Ideal Final Result

IFR: The system achieves [desired function] using [existing resource],
without adding [cost/complexity/harm], where [problem element] solves itself.

3. Inventive concepts (3–5)

For each concept:

• Concept name — short descriptive title
• TRIZ principle(s) applied — number + name (from resources/40_principles.md)
• Source — one of matrix (cell-derived), inferred (reasoning over the 40 principles), standard-solution- (Su-Field, see resources/76_standard_solutions.md), or separation- (physical contradiction, see resources/separation_principles.md)
• Description — specific, actionable engineering solution (2–3 sentences)
• Resource utilized — which internal / external / temporal resource is leveraged
• Implementation path — high-level technical pathway
• Risks / open questions — what would need to be validated experimentally

4. Principle-to-concept summary table

Sorted by Ideality descending.

| # | Concept | Principle(s) | Source | Key resource | Ideality | Risk | |---|---------|--------------|--------|--------------|----------|------|

5. Recommendation

State the top concept, the rationale (one sentence on ideality + risk), the next concrete validation action (experiment / prototype / simulation), and an escalation rule: if no concept reaches ideality > 1, invoke ARIZ (resources/ariz_85c.md).

Ideality metric

For each concept compute:

ideality = Σ (useful functions) / (Σ (harmful functions) + Σ (costs))

• Useful functions: each output the concept delivers (deceleration, heat dissipation, safety isolation, …) — count and, where possible, quantify.
• Harmful functions: side-effects the concept introduces (NVH, mass penalty, leaked field, …).
• Costs: capex, complexity, manufacturing penalty, certification burden, lifetime degradation.

Numeric estimation is preferred. When quantities are unavailable, use the bands low (≤1) | medium (1–3) | high (>3) and explain why a numeric estimate is impossible. A concept with ideality ≤ 1 must be dropped — TRIZ forbids compromise solutions.

Intake prompt

If any required input is missing, ask the user:

To run TRIZ analysis I need:

[ ] System description — what is the system and its primary function?
[ ] Problem statement — what bottleneck or unwanted effect must be resolved?
[ ] Contradiction — which parameter must improve, and which one degrades?
[ ] Constraints — cost, dimensional, environmental, manufacturing limits?
[ ] Resources — substances, fields, geometry, byproducts available?

Please fill the gaps so I can generate inventive concepts.

Agent instructions

1. Scope check — before intake, confirm the problem is engineering: physical system, physical field, measurable physical parameters. If not, refuse-with-reframe per examples/anti_example_misframed.md.
2. Intake — collect all five inputs; do not proceed with placeholders.
3. Analysis — execute Steps 1–5 in order. Step 6 (ARIZ-85C, resources/ariz_85c.md) and Step 7 (Trends, resources/evolution_trends.md) only on explicit request or as the ARIZ escalation triggered by failing the ideality bar.
4. Synthesis — produce 3–5 concrete concepts, not generic advice. Use resources/output_template.md verbatim.
5. Citation — always reference principle by #NN — Name and mark Source: matrix | inferred | standard-solution- | separation-.
6. Validation — every concept must (a) name the principle, (b) name the resource it consumes, (c) report a numeric or banded ideality estimate, (d) state at least one open validation risk.
7. Ideality bar — drop any concept with ideality ≤ 1. If no concept clears the bar, escalate to ARIZ-85C rather than compromise.
8. Anti-pattern — never compromise between A and B; the point of TRIZ is to resolve the contradiction, not split the difference. Compromise candidates must be listed as Source: rejected for discipline.

Example applications

• examples/brake_disc.md — automotive brake disc (friction vs heat dissipation). Mechanical / thermal contradiction; canonical TRIZ recovery of the ventilated-disc industry solution from first principles.
• examples/battery_pack.md — EV battery pack (energy density vs thermal safety). Electromechanical, contemporary; demonstrates the physical contradiction path via separation-by-condition and the 76 Standard Solutions.
• examples/heat_exchanger_fouling.md — petrochemical heat exchanger (fouling vs pressure-drop). Process-industry contradiction; demonstrates the 30→22 harm-versus-energy-loss pattern, Su-Field Class 1.2 destruction routing, and the ideality drop rule (one concept eliminated).
• examples/anti_example_misframed.md — UX onboarding flow (out-of-scope). Demonstrates the refuse-with-reframe behaviour when the problem is not engineering.

Resources

• resources/39_parameters.md — the 39 engineering parameters with definitions
• resources/40_principles.md — the 40 inventive principles with sub-principles
• resources/contradiction_matrix.json — full Altshuller 39×39 matrix (1190 populated cells out of 1482 non-diagonal; the 292 empty cells reflect contradictions for which Altshuller's analysis surfaced no dominant principle)
• resources/separation_principles.md — four-axis decision procedure for physical contradictions, linked to the 40 principles
• resources/76_standard_solutions.md — full 5-class / subclass taxonomy of the 76 Standard Solutions with Su-Field algebra notation, diagnostic flow, and rule texts (reconciled from Salamatov 1999, Mann 2002, ICG Training & Consulting materials)
• resources/ariz_85c.md — ARIZ-85C nine-part deep-analysis procedure with operational checklist
• resources/evolution_trends.md — eight trends of engineering system evolution + S-curve framing for roadmapping
• resources/glossary.md — operational definitions of every term used in the skill
• resources/output_template.md — machine-readable output template (use verbatim)

References

• Altshuller, G. (1984). Creativity as an Exact Science.
• Altshuller, G. (1999). The Innovation Algorithm: TRIZ, Systematic Innovation and Technical Creativity.
• Mann, D. (2002). Hands-On Systematic Innovation.
• Souchkov, V. Breakthrough Thinking with TRIZ for Business and Management.

Tags

#triz #engineering #problem-solving #innovation #systematic-invention #contradiction-resolution #inventive-principles

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: Antropocosmist
• Source: Antropocosmist/triz-engineering-solver
• License: MIT

Install and usage instructions live in the source repository linked above.