Hangchi's Manufacturing Software Pain Points & OntoL Ontology Solutions

Hangchi (Hangzhou Qianjin Gearbox Group) is a discrete manufacturer of heavy equipment — marine gearboxes and industrial transmissions — characterized by complex product structures, multi-version BOMs, frequent engineering change notices (ECNs), and long-term coexistence of multiple heterogeneous software systems. The article details six core pain points encountered during software implementation and maps each to OntoL's ontology-driven knowledge graph solutions.

1. Heterogeneous System Data Silos: PLM/ERP/MES Lack Unified Business Semantics

Hangchi deployed CAD, CAE, PLM, ERP, and MES from different vendors. These systems lack a unified business semantics layer, relying only on file export/import and point-to-point interface synchronization, which does not achieve true semantic integration.

PLM outputs BOMs to ERP, but material codes, versions, and units of measure often mismatch; drawing changes do not automatically propagate BOM versions to ERP/MES.

Design, process planning, and manufacturing interact serially via file transfers; models and simulation data cannot be automatically pushed to the shop floor.

The same material has different definitions in PLM, ERP, and MES, causing inconsistent inventory and work order statuses across systems, making business-finance reconciliation extremely difficult.

Traditional integration uses API interfaces for data synchronization, but interfaces only sync fields and cannot understand business semantics . When BOM versions, ECN changes, or WIP statuses change, interfaces merely sync static values without automatically reasoning about impact scope. OntoL's approach: establish a unified TBox ontology, mapping all PLM/ERP/MES data to a single business knowledge graph where materials, BOMs, versions, and ECNs share unified semantic definitions — not just simple interface docking.

2. Multi-Level BOMs and Frequent ECNs Cause Version Management Chaos

Gearbox products have deep hierarchies (assembly → sub-assembly → part). A single part drawing change ripples up to the top-level assembly BOM.

After an ECN, manual impact analysis is required to identify which BOMs, in-process work orders, purchase orders, and shop floor WIP materials need version switching — huge manual effort with high risk of omissions.

New and old version materials coexist in inventory; shop floor WIP mixes both versions. Systems only track quantities, not which BOM version each WIP item belongs to.

Common scenario: drawings are revised but the shop floor still issues old-version materials for production; version mismatch is discovered only after final assembly, causing rework and scrap.

Corresponds to the business-finance pain point: shop floor WIP contains large amounts of old-version materials; traditional ERP charges material issuance directly to cost. After scrap, tracing cost and attributing it to the correct work order is very difficult. OntoL solution: ontology axioms bind material versions and BOM effective intervals; ECN triggers automatic reasoning to identify affected BOMs, work orders, and WIP inventory, while preserving WIP cost attribution in the finance ontology layer.

3. Massive WIP in Long-Cycle Production Creates Business-Finance Disconnect and Cost Distortion

Heavy gearbox production cycles are long, spanning months or quarters. Large quantities of materials are issued to the shop floor and remain as WIP (castings, semi-finished, sub-assemblies) for extended periods.

Traditional ERP hard-binds documents: material issuance immediately accumulates cost, but physical items remain in WIP while work orders stay open; WIP value cannot be managed independently.

Business view: warehouse shows outbound; finance view: materials remain inventory assets and cannot be directly expensed — physical and financial views are disconnected .

Month-end closing requires manual shop floor WIP counting and manual WIP amount estimation — high effort, large estimation errors, severe monthly profit fluctuations.

High-cost gear blanks and forgings: when shop floor scrap occurs, tracing the original material batch, work order, and cost is difficult.

4. Master Data Governance Difficulty: High Maintenance Cost for Materials and Routings

Massive part count: gears, shafts, bearings, seals — tens of thousands of materials; historical coding inconsistent, one material multiple codes, one code multiple materials.

Complex routings: multi-operation, heat treatment, machining, assembly; different product versions have different routings; manual maintenance in traditional systems leads to errors on change.

Poor master data quality causes inaccurate MRP runs, failed kitting checks, unreliable shortage alerts, and production stoppages due to missing parts.

5. High-Mix Low-Volume Production: Planning and Shop Floor Execution Disconnect

Marine gearboxes follow Engineer-to-Order (ETO) + Make-to-Order (MTO) hybrid mode with strong order differentiation and frequent custom modifications.

ERP master production scheduling and MES shop floor execution are disconnected; ERP plans cannot account for shop floor equipment load, existing work orders, or WIP occupancy.

Weak reserved inventory control: sales orders and production orders reserve inventory, but traditional systems show available stock that is already reserved by other orders, preventing shipment or issuance.

6. Broken Change and Quality Traceability Chains

Gearboxes are high-end equipment requiring full traceability after delivery: finished assembly → part batch → blank heat number → raw material batch. Under traditional multi-system architecture, traceability requires cross-system multi-table SQL joins across PLM, ERP, MES — chains easily break. When quality issues arise, quickly identifying which work orders and which finished batches used the defective parts is difficult.

Pain Point to OntoL Solution Mapping

Multi-system silos, inconsistent PLM/ERP/MES data definitions → Unified TBox business ontology; heterogeneous system data mapped to single knowledge graph; unified semantics for materials, BOMs, work orders; no longer reliant solely on point-to-point interfaces.

Manual ECN impact analysis, high risk of missed/incorrect changes → Ontology layer defines BOM version effective axioms; ECN changes trigger automatic reasoning to identify affected BOMs, work orders, purchase orders, shop floor WIP.

Long-cycle production, large WIP, business-finance disconnect → Business ontology manages physical WIP state; finance ontology independently applies accounting standards — material issuance only accumulates WIP inventory; cost of goods sold recognized only upon completion + sales fulfillment; automatic reconciliation of physical vs. book differences.

WIP mixes new/old material versions, scrap traceability difficult → WIP as independent business object bound to material version, batch, work order, cost; one-click traceability on scrap; asset loss accounted separately.

Inaccurate MRP/kitting, weak reserved inventory control → Ontology automatically distinguishes available/reserved/frozen/WIP inventory; reasoning engine performs kitting validation and inventory availability judgment.

Cross-system quality traceability chain broken → Knowledge graph enables full-chain semantic traceability: finished product → parts → raw material batches integrated; no complex SQL joins needed.