产品说明书 · Product Specification Sheet
三层资产建模 · 实时状态追踪 · 三种仿真引擎，服务离散、流程、批次三类制造模式。
Three-tier asset modeling, real-time state tracking, and three simulation engines — covering discrete, continuous-process, and batch manufacturing.
数字孪生系统是道纳科技智能制造 AI 工厂平台中的独立微服务模块，为离散制造、流程制造与批次制造企业提供设备级—产线级—工厂级三层资产建模、实时状态追踪与多场景仿真能力，帮助企业在虚拟环境中验证产能规划、排产策略与异常影响，降低现场试错成本。
The Digital Twin System is an independent microservice within Daona Tech's AI Factory Platform. It provides three-tier asset modeling (machine → production line → factory), real-time state tracking, and multi-scenario simulation for discrete, continuous-process, and batch manufacturers — allowing capacity plans, scheduling strategies, and disruption impact to be validated virtually before committing to the shop floor.
产品定位面向大湾区制造业、贸易与产业投资服务客户：既可作为独立评估工具部署于试点产线，也可与道纳 AI 工厂平台其他模块（数据采集、边缘计算、AI 排产等）组合交付。
Positioned for Greater Bay Area manufacturing, trading, and industrial-investment clients: deployable standalone for a pilot line, or combined with other AI Factory modules (data collection, edge computing, AI scheduling).
以图结构组织资产层级：工厂 → 产线 → 设备，支持任意深度父子关系与子树查询，节点类型、元数据与属性均可自定义扩展。
Assets are organized as a hierarchical graph — factory → line → machine — supporting arbitrary-depth parent-child relations and subtree queries. Node type, metadata, and properties are all extensible.
每个节点属性均带时间戳与质量标记（good/bad），可通过 API 实时写入与查询，为看板、告警与仿真提供统一数据基础。
Every property value carries a timestamp and quality flag (good/bad), read/written via API in real time — a single consistent data foundation for dashboards, alerting, and simulation.
离散事件、连续流程、批次/整备三种可切换引擎，覆盖三类典型制造模式，统一输出 KPI 与事件日志。
Discrete-event, continuous-flow, and batch/campaign engines — interchangeable, covering three manufacturing modes, with a unified KPI and event-log output.
与平台内化工（GMP、联锁保护、配方生命周期）、纺织（浴比、色差、返修、缸类排产）组件共享数据模型，仿真结果可结合工艺约束校准。
Shares its data model with the platform's chemical (GMP, interlock safety, recipe lifecycle) and textile (liquor ratio, color difference, rework, dye-batch scheduling) components, so results can be calibrated against real process constraints.
| 引擎Engine | 适用模式Mode | 建模内容What It Models | 典型行业Typical Industries |
|---|---|---|---|
| 离散事件仿真Discrete Event Simulation | 离散加工装配Discrete manufacturing | 设备利用率、随机故障停机、批量生产与不良率Utilization, random breakdowns, batch output, defect rate | 精密加工、焊接装配、注塑Machining, welding assembly, injection molding |
| 连续流程仿真Continuous Flow Simulation | 连续工艺流程Process manufacturing | 流量波动、储罐液位、管道通量、能耗、断流事件Flow rate, tank level, throughput, energy use, interruptions | 化工、炼化Chemical, petrochemical |
| 批次/整备仿真Batch / Campaign Simulation | 批次生产Batch manufacturing | 批次周期、换品清洗、整备排产、OEE、批次合格率Batch cycle, changeover cleaning, campaign scheduling, OEE, batch yield | 制药、精细化工、纺织染整Pharma, fine chemicals, textile dyeing |
| 引擎Engine | KPI 输出字段KPI Fields | 事件类型Event Types | 默认参数Default Parameters |
|---|---|---|---|
| 离散事件仿真Discrete Event | 总产量、总不良数、良率、平均利用率、故障次数total_produced, total_defects, yield_rate, avg_utilization, breakdowns | 故障停机（恢复时长）、质量问题（不良数/产出数）breakdown (recovery time), quality_issue (defects/produced) | 时长默认 480 分钟（8 小时班次），步长默认 60 秒Duration 480 min default (8-hr shift), step 60s default |
| 连续流程仿真Continuous Flow | 总通量、总损耗、良率、断流次数、能耗total_throughput, total_waste, yield_rate, flow_interruptions, energy_consumed | 断流（恢复时长）、液位告警（液位百分比）flow_interruption (recovery time), tank_level_alert (level %) | 同上，可按场景自定义Same defaults, scenario-customizable |
| 批次/整备仿真Batch / Campaign | 总批次、合格批次、不合格批次、批次合格率、清洗次数、清洗总时长、设备综合效率total_batches, passed_batches, failed_batches, batch_yield, cleaning_events, total_cleaning_min, oee | 换品清洗（源/目标产品/时长）、批次失败（产品/批次号/所属整备）cleaning (from/to/duration), batch_failure (product/batch/campaign) | 同上，可按场景自定义Same defaults, scenario-customizable |
场景名称、仿真时长与步长均可在发起仿真时自定义传参，便于对比不同排产策略或异常场景的影响。Scenario name, duration, and step interval are all configurable per run, making it easy to compare scheduling strategies or disruption scenarios.
化工模块 Chemical Module
纺织模块 Textile Module
数控加工、焊接装配、注塑成型产线的设备利用率与不良率评估。
Utilization and defect-rate evaluation for CNC machining, welding-assembly, and injection-molding lines.
化工、炼化连续流程的储罐液位、管道通量与能耗仿真。
Tank-level, pipeline-throughput, and energy simulation for chemical and petrochemical processes.
制药、精细化工、纺织染整的换品清洗、整备排产与批次合格率仿真。
Changeover, campaign-scheduling, and batch-yield simulation for pharma, fine chemicals, and textile dyeing.
虚拟环境中提前验证扩产、排产调整或设备升级方案的产出与风险。
Validate expansion, rescheduling, or upgrade plans virtually before committing capital or shop-floor time.
系统内置示例工厂涵盖离散、连续流程与批次三类产线（产线A-精密加工、产线B-焊接装配、产线C-注塑成型），可直接作为演示基线，亦可替换为客户真实资产结构。The built-in demo factory spans discrete, continuous, and batch line types (Line A-precision machining, Line B-welding assembly, Line C-injection molding) and can serve as a demo baseline or be replaced with a client's real asset structure.
| 行业Industry | 支持深度Support Level | 对应能力Capability Used | 典型应用Typical Application |
|---|---|---|---|
| 精细化工Fine Chemicals | ★★★ 深度支持（专属组件）Dedicated module | 批次/整备仿真 + 化工模块Batch/Campaign Simulation + Chemical Module | 反应釜产能与安全联锁验证、配方放大审批、批次合格率仿真Reactor capacity & interlock validation, recipe scale-up approval, batch-yield simulation |
| 制药Pharmaceutical | ★★★ 深度支持（复用化工模块合规组件）Dedicated (shares chemical module) | 批次/整备仿真 + 化工模块（21 CFR Part 11 电子签名、偏差CAPA）Batch/Campaign Simulation + Chemical Module | 电子批记录审计追溯、偏差调查全流程留痕、批次合格率仿真Electronic batch-record audit trail, end-to-end deviation tracking, batch-yield simulation |
| 纺织印染Textile Dyeing & Finishing | ★★★ 深度支持（专属组件）Dedicated module | 批次/整备仿真 + 纺织模块Batch/Campaign Simulation + Textile Module | 缸差排产优化、色差自动判定、下料清单自动计算Dye-batch scheduling optimization, automated color-difference verdict, automatic material dosing |
| 精密机械加工 / 汽车零部件Precision Machining / Auto Parts | ★★ 广泛适用（通用引擎）Broadly applicable | 离散事件仿真Discrete Event Simulation | 设备利用率与故障影响评估、不良率基线测算（对应示例产线A）Utilization & breakdown-impact evaluation, defect-rate baseline (mirrors demo Line A) |
| 电子电器组装Electronics & Appliance Assembly | ★★ 广泛适用（通用引擎）Broadly applicable | 离散事件仿真Discrete Event Simulation | 装配产线利用率与质量问题追踪（对应示例产线B焊接装配）Assembly-line utilization & quality tracking (mirrors demo Line B) |
| 塑料制品 / 注塑加工Injection Molding & Plastics | ★★ 广泛适用（通用引擎）Broadly applicable | 离散事件仿真Discrete Event Simulation | 批量生产与不良率仿真（对应示例产线C）Batch output & defect-rate simulation (mirrors demo Line C) |
| 炼化 / 石化下游Petrochemical & Refining Downstream | ★★ 广泛适用（通用引擎）Broadly applicable | 连续流程仿真Continuous Flow Simulation | 储罐液位、管道通量、能耗仿真Tank-level, pipeline-throughput, and energy simulation |
| 食品饮料连续化产线Food & Beverage (Continuous Lines) | ★ 概念可延伸Extensible in principle | 连续流程仿真（通用模型，尚无专属组件）Continuous Flow Simulation (generic, no dedicated module) | 灌装、巴氏杀菌等连续产线的流量与能耗仿真，需按客户工艺参数配置Flow/energy simulation for filling, pasteurization, etc., pending configuration to actual process parameters |
★★★ 行业（精细化工、制药、纺织印染）已具备与其工艺强相关的专属组件（详见"核心能力"第 4 项）；★★ 行业目前复用通用离散/连续仿真引擎，尚无该行业专属工艺组件；★ 行业仅停留在模型概念可延伸层面，需结合客户实际工艺参数定制开发。行业专属组件的扩展节奏可根据具体客户试点需求排期。
Industries marked ★★★ (fine chemicals, pharma, textile dyeing) already have dedicated process modules (see Core Capabilities item 4). Industries marked ★★ currently reuse the generic discrete/continuous engines with no dedicated vertical module yet. The ★ industry is extensible in principle only, pending custom development against a client's actual process parameters. Further vertical-module development can be scheduled against specific client pilots.
当前版本为可运行的独立仿真引擎服务，图数据采用进程内内存存储（服务重启后重置），面向概念验证（POC）与试点场景，尚未包含持久化存储、3D 可视化前端与实时数采链路的自动对接——这些为后续路线图内容，可根据客户试点需求排期。
The current version is a working, standalone simulation service. Twin-graph data is held in memory (reset on service restart), suited to proof-of-concept and pilot use. Persistent storage, a 3D visualization front-end, and automated real-time data-ingestion integration are not yet included — these are roadmap items, schedulable against a client's pilot requirements.
在虚拟环境中验证产能规划与排产策略，缩短方案验证周期，降低现场试错成本。
Validate capacity and scheduling plans virtually — shortening the validation cycle and reducing shop-floor trial-and-error cost.
三种仿真引擎覆盖离散、流程、批次三类典型制造模式，适配大湾区制造业的多元工艺结构。
Three simulation engines cover discrete, process, and batch manufacturing — matching the diverse process mix across the Greater Bay Area's manufacturing base.
作为道纳 AI 工厂平台的独立模块，可单独试点，亦可与数据采集、AI 排产、预测性维护等模块组合，按客户节奏分期交付。
As a standalone AI Factory module, it can be piloted alone or combined with data-collection, AI-scheduling, and predictive-maintenance modules, phased to the client's own pace.