Forge Industrial Division · Product Lifecycle Management

Platform Technical
Design Document

Architecture, knowledge graph design, engine specification, and cross-engine integration across six specialized PLM engines — requirements traceability, BOM configuration, engineering change management, simulation orchestration, regulatory compliance, and field feedback intelligence — competing directly with Siemens Teamcenter, PTC Windchill, and Dassault 3DEXPERIENCE. Built in Rust.

Engines
6 PLM Intelligence Systems
Core Runtime
Rust · Zero-Allocation Hot Paths
Impact Analysis
<10s Across 50K+ Components
Classification
Confidential — ITAR Controlled
Architecture
Six Engines. One Knowledge Graph.
Each engine is a standalone sub-platform. Together, they form the digital thread.
MERIDIAN
Requirements Traceability
MBSE · V&V matrix · DO-178C coverage · Bi-directional trace from need to test
LATTICE
Multi-View BOM Configuration
eBOM/mBOM/sBOM/aBOM · Variant config · CAD sync · As-built serial trace
CASCADE
Engineering Change Management
Graph impact analysis · Parallel approval · Effectivity · Hunt-to-detection flywheel
NEXUS
Simulation Integration
FEA/CFD/MBD orchestration · Generative design · Digital twin validation
SENTINEL
Regulatory Compliance
21 CFR Part 11 · AS9100 · ITAR/EAR · ALCOA+ audit · E-signatures
ECHO
Field Feedback Intelligence
Warranty analytics · CAPA-to-ECR · Reliability prediction · Closed-loop design
Executive Summary
Unified Architecture
Forge Axiom is a Rust-built product lifecycle management platform that competes directly with Siemens Teamcenter (the ABI Research and Forrester Wave 2025 leader), PTC Windchill, and Dassault 3DEXPERIENCE. The architectural differentiation is fundamental: where incumbent PLM vendors bolt digital thread capabilities onto relational database architectures designed in the 1990s, Axiom is built from the ground up on a product knowledge graph — a native graph data structure where every requirement, design artifact, BOM component, simulation result, change order, compliance record, and field failure is a node, and every relationship between them is an edge. This architecture enables sub-10-second full-scope impact analysis across 50,000+ component products (versus minutes to hours in traditional PLM), native digital thread without middleware integration, and cross-engine intelligence flows that are architecturally impossible in siloed relational systems.
The platform comprises six specialized engines, each a standalone sub-platform with its own deep technical architecture, connected through the shared knowledge graph. Meridian (cardinal red) manages requirements traceability with MBSE and bi-directional V&V coverage. Lattice (structural teal) manages multi-view BOM configuration with four simultaneous views (eBOM/mBOM/sBOM/aBOM) resolving millions of variant configurations in milliseconds. Cascade (electric blue) manages engineering change with graph-based BFS impact analysis on Rust zero-allocation hot paths. Nexus (simulation violet) orchestrates FEA/CFD/MBD simulation workflows with generative design integration. Sentinel (compliance white-gold) enforces 21 CFR Part 11, AS9100, ITAR/EAR with ALCOA+ audit trails and write-path e-signature interception. Echo (signal amber) closes the loop from field failures back to root-cause design decisions, completing the digital thread from requirement to retirement.
MERIDIAN
Requirements & MBSE
LATTICE
BOM Configuration
CASCADE
Change Management
NEXUS
Simulation & DT
SENTINEL
Compliance & Audit
ECHO
Field Intelligence
Meridian
Requirements Traceability & MBSE
ENGINE 01 — MERIDIAN
Requirements Traceability & Systems Engineering
Every requirement traced from stakeholder need through design, verification, and validation — bi-directionally

Meridian manages the foundational layer of the digital thread: the requirements that define what the product must do. The engine implements bi-directional traceability from stakeholder needs through system requirements, subsystem requirements, design specifications, verification procedures, and validation evidence. Every requirement is a node in the knowledge graph, linked to the design artifacts that satisfy it (Lattice BOM components, CAD models), the simulations that verify it (Nexus), the test procedures that validate it, and the change orders that modify it (Cascade). Coverage analysis computes V&V completeness in real time — "requirement REQ-4207 is satisfied by design artifact DA-8402, verified by simulation SIM-1204 (pass, 2024-11-15), but validation test VAL-4207 has not been executed." DO-178C structural coverage analysis for avionics, ISO 26262 ASIL decomposition for automotive, and IEC 62304 software safety classification for medical devices are implemented as domain-specific requirement classification schemas.

100%
Bi-directional traceability from need to test evidence
DO-178C
Structural coverage analysis for avionics safety levels A-E
Real-time
V&V coverage gap detection as requirements and designs evolve
Lattice
Multi-View BOM Configuration
ENGINE 02 — LATTICE
Multi-View BOM & Variant Configuration
Four views. One truth. Zero drift. Millions of configurations resolved in milliseconds.

Lattice manages the most critical and most poorly managed data structure in all of manufacturing: the Bill of Materials. The engine maintains four simultaneous BOM views in a single graph — eBOM (engineering, as-designed), mBOM (manufacturing, as-planned with process materials and routing), sBOM (service, as-maintained with field-replaceable units and supersession chains), and aBOM (as-built, serialized per-unit configuration captured at each manufacturing station). Cross-view reconciliation runs continuously, detecting discrepancies between engineering design intent and manufacturing reality in real time. The variant configuration engine resolves a 150% super-BOM into customer-specific configured BOMs using constraint-satisfaction algorithms that process millions of valid configurations in milliseconds — replacing the manual configuration spreadsheets that cause 40% of order-entry errors in engineer-to-order manufacturers. CAD-to-BOM synchronization extracts BOM structures natively from CATIA, NX, Creo, SolidWorks, Inventor, Fusion 360, and STEP/JT neutral formats.

4-view
Simultaneous eBOM/mBOM/sBOM/aBOM in single graph structure
ms
Variant configuration resolution for millions of valid combinations
99.7%
BOM accuracy after CAD synchronization (vs. 85% industry average)
Cascade
Engineering Change Intelligence
ENGINE 03 — CASCADE
Engineering Change Management & Impact Analysis
Sub-10-second full-scope impact analysis across 50,000+ component products — on the knowledge graph

Cascade manages the mechanism by which the product definition evolves: engineering changes. The core architectural innovation is graph-based impact analysis using modified BFS traversal on the Axiom product knowledge graph, replacing the recursive SQL queries against relational tables that traditional PLM systems use (an approach that degrades exponentially with product complexity). When an engineer proposes changing a material from aluminum 6061-T6 to titanium Ti-6Al-4V for a bracket assembly, Cascade's impact graph traverses every relationship from that component outward: parent assemblies affected, simulation models that must be re-verified (Nexus), manufacturing routings that must be updated, supplier AVL changes required, cost implications across all BOM views (Lattice), and regulatory submissions that must be amended (Sentinel). The BFS traversal runs on Rust zero-allocation hot paths with hash-indexed adjacency lists, achieving sub-10-second full-scope impact analysis across products with 50,000+ components. Parallel approval orchestration uses DAG-based dependency resolution to route ECRs through concurrent review gates.

<10s
Full-scope impact analysis on 50K+ component products
O(1)
Edge lookup via hash-indexed adjacency lists (vs. O(n) SQL joins)
95%
First-pass yield in mature change environments using Cascade
Nexus
Simulation & Digital Twin
ENGINE 04 — NEXUS
Simulation Integration & Generative Design
Every simulation versioned. Every result traced. Every design decision defensible.

Nexus orchestrates simulation workflows across FEA (Ansys, Abaqus, Nastran), CFD (Fluent, OpenFOAM, Star-CCM+), multi-body dynamics (Adams, RecurDyn), and electromagnetic simulation (CST, HFSS) — managing simulation data as first-class citizens in the knowledge graph rather than file attachments in a PDM vault. Every simulation model, mesh, boundary condition set, solver configuration, and result dataset is a node linked to the design artifact it validates, the requirement it verifies (Meridian), and the engineering change that triggered it (Cascade). The simulation data management architecture solves the critical problem that incumbent PLM vendors handle poorly: when a design changes, which simulations need to be re-run? Nexus answers this automatically through the knowledge graph — a material change in Cascade triggers re-verification of every simulation whose results depend on material properties. Generative design integration enables constraint-driven topology optimization with TPMS lattice generation for additive manufacturing.

Auto
Re-simulation triggering when design changes affect verified parameters
6+
Solver integrations (Ansys, Abaqus, OpenFOAM, Adams, CST, HFSS)
Graph
Simulation data as knowledge graph nodes, not file attachments
Sentinel
Regulatory Compliance & Certification
ENGINE 05 — SENTINEL
Regulatory Compliance & E-Signature Architecture
Compliance is not a documentation exercise — it is an architecture feature

Sentinel transforms regulatory compliance from a post-hoc documentation exercise into an architectural feature of the PLM system itself. The write-path interceptor architecture ensures that every data mutation across all Axiom engines passes through Sentinel's compliance layer before persistence — capturing who made the change, when, why, what the previous state was, what the new state is, and preserving cryptographic proof that the record has not been altered since capture. FDA 21 CFR Part 11 compliance is achieved through ALCOA+ audit trails (Attributable, Legible, Contemporaneous, Original, Accurate + Complete, Consistent, Enduring, Available) with Part 11-compliant electronic signatures requiring biometric or two-factor authentication, signature meaning declaration (authored, reviewed, approved), and immutable binding between the signature and the specific data version signed. ITAR/EAR export control enforces per-request access evaluation against USML category classifications with Technical Assistance Agreement lifecycle management.

ALCOA+
Eight-attribute audit trail compliance for FDA/EMA submissions
Part 11
21 CFR Part 11 e-signature with biometric/2FA and meaning declaration
ITAR
Per-request export control with USML category classification
Echo
Field Feedback & Closed-Loop Design
ENGINE 06 — ECHO
Field Feedback Intelligence & CAPA Integration
Closing the loop from field failure to root-cause design decision — completing the digital thread

Echo closes the digital thread by connecting field performance data — warranty claims, service reports, IoT sensor telemetry, customer complaints, and recall events — back to the design decisions that caused them. When a field failure is reported, Echo traces the failure through the as-built BOM (Lattice aBOM) to identify the specific component, revision, and supplier lot, then traverses the knowledge graph to find the design decision, simulation result, and requirement that governed the failed parameter. The CAPA-to-ECR pipeline automates the corrective action workflow: a validated root cause with design-related attribution auto-generates a draft Engineering Change Request in Cascade, pre-populated with failure evidence, affected population estimate (from the aBOM serial registry), and proposed corrective action. Weibull reliability prediction models estimate time-to-failure distributions for components and assemblies, enabling proactive design changes before field failure rates exceed warranty cost thresholds. This closed-loop architecture is what traditional PLM systems cannot achieve because they store field data in separate quality management systems disconnected from the product knowledge graph.

Closed
Loop from field failure to root-cause design decision via knowledge graph
Auto
CAPA-to-ECR pipeline with pre-populated failure evidence and affected population
Weibull
Reliability prediction for proactive design changes before warranty threshold
Cross-Engine Intelligence
The Knowledge Graph Connects Everything
Meridian → Cascade: Requirement change triggers impact analysis across all implementing design artifacts. Re-verification scope auto-calculated. Draft ECR generated with requirement change as justification.
Lattice → Cascade: BOM revision propagated atomically across all four views. Effectivity state machine manages configuration cutover across manufacturing lines.
Cascade → Nexus: Design changes affecting simulation-verified parameters trigger re-simulation workflow. Verification matrix updated when new results arrive.
Sentinel → All Engines: Every data mutation across all engines passes through the write-path interceptor. E-signatures, audit trails, and export controls enforced at the persistence layer.
Echo → Cascade: Field failures with design-related root causes auto-generate draft ECRs with failure evidence, affected population, and proposed corrective action. CAPA-to-ECR loop closed.
Nexus → Lattice: Generative design outputs create new BOM items with manufacturing feasibility tags. Simulation-verified material properties propagate to BOM cost models.