Smart Factory and Industrial IoT: Making Production Line Data Visible, Decidable, and Actionable in Real Time

The U.S. National Institute of Standards and Technology (NIST) defines smart manufacturing as "a fully integrated, collaborative manufacturing system that responds in real time to changing demands and conditions in the factory, in the supply network, and in customer needs." In software terms, that means five system layers, one data pipeline, and one table of metric definitions. Jiding Technology delivers exactly this for manufacturers in Suzhou and the Yangtze River Delta: from auditing equipment data points and integrating protocols, through MES, shop-floor dashboards, and an IoT platform. On the equipment side we handle software and integration only; we do not take on hardware general contracting or subcontract it out.

A Smart Factory Isn't a Software Purchase — It Grows Across Five Layers

Most failed projects treat the five layers as a one-time purchase. What actually works is making each layer deliver value on its own and fund the next layer by itself.

① Decision & Management LayerProduction Planning & Scheduling (APS)OEE & Yield AnalysisOrder Delivery ForecastingEnergy & Cost AllocationMobile Approvals② Manufacturing Execution LayerWork Orders & RoutingProduction Reporting & WIPQuality & SPCBatch TraceabilityLine-side Storage & Material Handling③ IoT Platform LayerDevice Onboarding & Thing ModelRule Engine & AlarmsTime-series Data StorageEdge Computing & Offline ResendOpen APIs & Subscriptions④ Acquisition & Network LayerIndustrial Gateway / DTUOPC UA · ModbusMQTT ReportingIndustrial Switches & Wi-Fi4G / 5G Backup⑤ Equipment & Field LayerPLC / CNCRobots / Special-purpose MachinesSensors & InstrumentsBarcode Scanners / RFIDVision & Weighing↑ Data flows bottom-up: milliseconds → seconds → minutes · ↓ Commands flow top-down: work orders, recipes, parameter downloads
The five-layer structure of a smart factory. The highlighted IoT platform layer is the one most often skipped: without a unified thing model, every new device means another code change, and by the third device the project stalls.

We Deliver Software and Integration

Gateways, sensors, PLCs, and switches are supplied by you or your integrator. We provide the selection list and acceptance criteria, and we turn their signals into usable data. We don't mark up hardware, and we're not a cabling crew that disappears after installation.

Phase One Commits to Just Three Numbers

On-time production reporting rate, visibility of equipment utilization, and traceability query time. We measure a baseline for all three before signing, then measure again at acceptance. If a client can't state a baseline and won't measure it, we don't take the project.

Four Roles Must Be Present

The production manager defines the metrics, the equipment supervisor defines the data points, quality defines the acceptance criteria, and IT defines the network and accounts. If one is missing, the project stalls on that person — the most common cause of delay, and almost never a technical one.

Five Module Types: Which One Goes First Matters More Than the Technology

The five modules can be delivered independently or stacked. The table below shows typical scope, duration, and first-phase budget ranges, so you can judge how big phase one should be.

ModuleProblem it solvesCore functionsTypical durationFirst-phase budget
MES / Manufacturing Execution Where each order stands, which operation is the bottleneck, how much WIP there is Work order dispatching, operation routing, barcode reporting, WIP dashboards, quality disposition and rework 8–14 weeks CNY 300k–800k
Equipment Connectivity & Data Acquisition Utilization estimated by hand, failures reported by shouting, downtime reasons never recorded Gateway acquisition, state detection, alarm push, automatic OEE calculation, equipment registry 4–8 weeks CNY 100k–300k
Shop-floor Dashboards Data lives in the shift leader's head, and management can't see real progress Output and takt time, exceptions and downtime, order progress, multi-screen rotation and role-based views 3–6 weeks CNY 50k–150k
IoT Platform Devices are scattered across plants and customers, and every new device type means another code change Thing model, device onboarding and authentication, rule engine, OTA, multi-tenancy and open APIs 8–16 weeks CNY 200k–600k
Automation Control Critical process parameters drift, and veteran operators adjust them by hand Closed-loop tuning, thresholds and interlocking, parameter downloads and auditing, exception and recovery logging Per project CNY 150k–600k
Three types of projects we don't take: hardware-only requests with no software (installing sensors without closing the data loop); "build it first, then we'll discuss requirements" (uncontrolled scope, guaranteed rework); and promises to "raise OEE by 20%" without measuring a baseline (no baseline means no acceptance criteria, and the project ends in disputes).

From Equipment to Dashboard, a Single Data Point Passes Through Six Stages

When a dashboard doesn't work, the problem is almost never the dashboard itself but the five stages before it. Each stage has a typical latency — and a typical pitfall.

01EquipmentPLC · MetersSensors · Barcode50 ms02Industrial GatewayProtocol conversionPoint acquisition1 s03Edge NodeCleaning · CachingOffline resend5 s04PlatformTime-series DBRule engine · Alarms2 s05Business SystemsMES · Work ordersBatch traceability30 s06PresentationShop-floor dashboardMobileIs the offline scenario covered?Edge nodes must cache at least 7 days of raw data locally, resend it by timestamp once the network recovers, and deduplicate idempotently in the cloud. Otherwise one shop-floor outage leaves a permanent gap in that batch of data.
The complete path of a data point from equipment to dashboard, with typical latency at each hop. Latency isn't technical showing off: a dashboard that lags more than 3 seconds never gets a second look from the shift leader.

Four Hard Rules for Data Quality

If these four rules aren't written into the design document, the system will need rework three months after go-live. They have nothing to do with technology — only with agreements.

  • Define point naming rules first: line–station–machine–measurement–unit, e.g. L2-ST03-WELD01-CURR-A. Get the names wrong and every report downstream has to change with them.
  • State detection must have evidence: running / standby / fault / stopped is determined by PLC signals, not guessed from current thresholds — nobody trusts a guessed utilization rate.
  • Timestamps follow the equipment: gateways sync with the server periodically and log clock drift. All cross-shift and cross-time-zone statistics depend on this.
  • Offline operation needs a fallback: edge nodes cache at least 7 days of raw data locally, resend by timestamp after recovery, and deduplicate idempotently in the cloud. One shop-floor outage must not leave a permanent gap in that data.
Automated production line: equipment modules, status indicator lights, and a shop-floor data dashboard
Every module on an automated line is a data source. What really decides whether a project succeeds is whether those signals are named, archived, and interpreted.

How to Connect Equipment: Nine Acquisition Methods and Their Real Costs

"It supports OPC UA" means ten different things at ten equipment vendors. This table is based on hands-on integration experience and is meant for estimating on-site effort.

Method / protocolWhat you can collectTypical equipmentDifficultiesIncremental cost
Modbus RTU (RS485)Register values, coil statesVFDs, temperature controllers, power meters, legacy PLCsThe vendor must provide the address table; the same register name means different things at different vendorsLow
Modbus TCPRegister values, coil statesInstruments with Ethernet ports, compact PLCsStation number and port configuration; access conflicts with multiple mastersLow
OPC UAStructured variables with types and semanticsSiemens, Beckhoff, and Rockwell current modelsThe vendor must expose a UA server and permissions; older models need a separate OPC serviceMedium
Siemens S7DB blocks, M area, I/Q pointsS7-200 / 300 / 400 / 1200 / 1500DB offsets must be computed from the data structure; symbol tables don't match across versionsMedium
Mitsubishi MC / Omron FINSData registers, link areasJapanese PLCs and machine toolsLow serial speeds on older equipment limit the acquisition cycle to 200 ms or moreMedium
MQTTPackets actively pushed by devices or gatewaysModern instruments, DTUs, AGVs, energy metersTopic design, QoS levels, offline message buildupLow
Vendor-proprietary serial protocolsMessage frames parsed per vendor documentationWeighing indicators, inspection devices, legacy special-purpose machinesDocumentation is often missing and must be reverse-engineered from packet captures; the timeline is hard to controlHigh
IO / Dry contactsOnly on/off and present/absentLegacy equipment, stack lights, safety doors, cylindersNo process values, only statesLow (and so is the information density)
Manual / Barcode entryTablet scanning at the station, manual data entryManual workstations, outsourced operationsTimeliness and accuracy depend on management, not technologyLow
Fill in one table before acquisition starts: equipment list × communication method × collectible points × acquisition cycle × responsible party (who provides documentation, who grants permissions). We provide this table during the free requirements assessment, and it is the only basis for deciding how far phase one can go. Book an equipment connectivity feasibility assessment →

OEE: Agree on the Definitions First, Then Talk About Improvement

On the same machine in the same month, a different definition can produce a 15-point difference. So the first step isn't buying a system — it's writing the definitions of the three metrics into a table and having both sides sign it.

Availability A90.2%563 / 624 hRun time ÷ planned production time×Performance P82.1%462 / 563 hOutput at theoretical cycle time×Quality Q95.0%439 / 462 hGood units ÷ total output×Overall Equipment Effectiveness (OEE)70.3%Industry reference: 55–75% for discrete manufacturing13%8%14%3%61%Planned downtime 96 hDowntime loss 61 hSpeed loss 101 hQuality loss 23 hGood output time 439 h
On the same machine in the same month, a different definition can produce a 15-point difference. How the 720 calendar hours are deducted, whether changeovers count as downtime, and whether trial-run parts count as scrap all have to go into the "Metric Definitions Table" before you talk about improvement.
MetricDefinitionWhere the data comes fromMost common mistake
Availability A Run time ÷ planned production time. Planned production time = calendar time − planned downtime (maintenance, no orders, holidays) PLC run signal + production schedule Counting planned downtime as loss keeps OEE low forever — and points improvement in the wrong direction
Performance P Output time at theoretical cycle time ÷ run time (including idling, minor stops, and slow running) Output count + standard process cycle time Whether theoretical cycle time uses the "historical best" or the "nameplate rating" must be fixed once and never swapped mid-project
Quality Q Good units ÷ total output. Total output includes trial-run parts and reworked parts Quality inspection results + rework records Trial-run parts quietly excluded make yield look high while the costs are real

Five Common Definition Pitfalls

  • Does changeover time count as downtime? Yes, but it must be tracked separately as "changeover loss." Mix it into failure downtime and you'll buy more expensive equipment, when what you actually need is SMED (single-minute exchange of dies).
  • How do you calculate a mixed-model line? Weight by output per product — a simple average lets low-volume, high-cycle-time products inflate overall performance.
  • Who fills in downtime reasons? Let operators pick from a list, never type free text. Three months later, free-text reasons exist in a hundred variations and none of them can be aggregated.
  • Do you split days by 24 hours or by shift? MES by shift and dashboards by calendar day is the number-one cause of "the two screens don't match."
  • Is an offline device counted as downtime? No, but track "acquisition unavailable time" separately. Otherwise a gateway dropping offline gets recorded as a machine failure, and the maintenance team makes a wasted trip.

The Core of an IoT Platform Isn't How Many Devices It Connects — It's the Thing Model Table in the Middle

The thing model is the single contract between devices and applications. Define it once and both sides can evolve independently: replace a device without touching the app, change the app without touching the device.

Device sideThing model: the single contract between devices and applicationsApplication sideIndustrial Gateway / DTUTelemetry PropertiesTemperature · Pressure · Speed · Count · CurrentShop-floor DashboardEdge Computing BoxEventsAlarms · Fault codes · Access control · ChangeoverMobile AppCameras & VisionServicesStart/stop · Recipe download · Remote restartWork Orders & AlarmsMeters & SensorsTagsLine · Station · Customer · ContractEnergy & CostPLC / RobotsShadowDesired state vs. reported state · consistent while offlineOpen APIDefine the thing model once, then devices and applications evolve independently: replace a device without touching the app, change the app without touching the device.
The real core of an IoT platform isn't how many devices it connects — it's the thing model table in the middle. It decides whether adding a production line three years from now takes two days of configuration or two months of development.
Thing model fieldExampleWhy it's required
Device IDGW01-PRESS-07Unique and immutable. Rename, relocate, or move a device to another plant, and its history stays attached to the same ID
Point keyspindle_tempSemantically stable and decoupled from register addresses. Swapping the PLC model only changes the mapping, not the reports
Data typefloat / int / boolDetermines the storage type and how values are rendered; get the type wrong by one bit and the curve turns into a sawtooth
Unit and dimension°C, mm/s, kWhWhether conversion happens on the front end or the back end must be decided at design time — never compute it in both places
Precision and dead band0.1 / dead band 0.5Directly drives reporting frequency and storage cost. Without a dead band, a single temperature curve can consume a year of storage
Acquisition cycle1 s / 60 s / event-triggeredCalculate cost backward from time-series database capacity; not every point needs second-level resolution
Ownership dimensionsLine / Station / Customer / ContractDetermines who can see what and how costs are allocated. A multi-tenant platform without this column can't charge external customers

Rules and Alarms

Thresholds, N consecutive breaches, debouncing, in-shift accumulation, and linked work orders. Alarms must reach a person: who takes it, how fast they must respond, and who it escalates to. Otherwise it's just a wall of red on the screen.

Time-series Data Service

Raw data is downsampled into minute, hour, and day tiers; raw values are retained for 1–2 years and aggregates indefinitely. Queries go through aggregate tables, so pulling up a curve from a year ago doesn't make you wait.

Device Management and OTA

Asset registry, lifecycle, remote configuration, firmware upgrades, and rollback. Upgrades must support staged rollout: one device first, then one line, then the whole plant, with automatic rollback on failure — skip this step and you can brick every gateway in the plant.

Edge or Cloud: Nine Out of Ten Factories Should Choose Hybrid

The edge handles what can't wait; the cloud handles what takes a long time to compute. Blur the two together and you either spend far too much or fall over the moment the network drops.

One Criterion Is Enough

If an action must complete within 100 milliseconds, or must keep running even when the network is down, it belongs on the edge. Everything else goes to the cloud. By this criterion, the boundary is never blurry.

  • Edge: acquisition and protocol conversion, state detection, offline caching and resending, real-time interlocking and audible/visual alarms, local HMI.
  • Cloud / data center: historical storage and aggregation, cross-line and cross-plant comparison, reporting and cost accounting, model training, multi-tenancy and open APIs.
  • Hard line: safety interlocks never go in the cloud. A design where a network hiccup stops the equipment isn't a smart factory — it's an accident waiting to happen.
IoT site diagram: devices and sensors linking up to the cloud through gateways over red data links
A typical topology: device → gateway → edge node → cloud. The red links in the diagram carry data upstream; control commands travel on a separate channel.
DimensionEdge onlyCloud onlyHybrid (recommended)
Acquisition latencyMilliseconds100 ms–2 s, affected by public network jitterMillisecond-level decisions + second-level upload to the cloud
Availability during outagesFully availableData interrupted, control lostSite keeps running, data is backfilled
History and cross-plant analyticsLimited capacity, hard to doStrongStrong
Cost per siteOne box per line, hardware cost up frontLow, pay as you goModerate, but can be phased
Best suited forStandalone machines, offline scenariosEnergy monitoring, remote operationsThe vast majority of manufacturing sites

Three Rules for Shop-floor Dashboards, and Four Typical Mistakes

A dashboard isn't "moving reports onto a wall." It serves exactly one scenario: a shift leader standing three meters away needs to know within ten seconds whether to call someone.

Rule 1: One Screen Answers One Question

Shift output versus target, equipment exceptions, or order progress — pick one per screen. Try to show all three at once and none of them will be legible.

Rule 2: Legible from Three Meters Away

Primary numbers at least 120 px, with no more than 7 numbers per screen. Everything else drops to secondary information or comes off the screen entirely.

Rule 3: Exceptions First

Normal states stay uncolored; only exceptions use red plus blinking plus sound, and must pinpoint the specific station and machine within 30 seconds. A screen covered in red, yellow, and green is the same as no signal at all.

Three Lines of Small Print You Must Include

Data as-of time, the statistical definition (including whether trial-run parts count), and the owner. Without these three lines, the floor starts doubting the numbers — and then stops looking at them.

Four Typical Mistakes

  • Cramming 20 charts onto one screen — more information means slower decisions. Cut it down to 7 numbers first, then talk about making it look good.
  • Using red, yellow, and green for three unrelated things — when red means "stopped," "over target," and "pending inspection" all at once, the floor simply ignores every color.
  • Five-minute data delay — a dashboard lagging more than 3 seconds never gets a second look from the shift leader; lag more than a minute and it only lights up when management visits.
  • Building one screen with no rotation or permissions — production wants output, the equipment team wants faults, the plant manager wants orders. Three rotating views are far more useful than one screen that shows "everything."

Five Steps, Each with Its Own Acceptance Criteria

Projects that skip step two and go straight to plant-wide MES usually end up eight months later with nothing but a dashboard nobody looks at. Here is the pace we actually use.

STAGE 01Inventory & Line Selection12–3 weeksEquipment list · Protocols · Collectible pointsAcquisition feasibility reportFree of chargeSTAGE 02Single-line Connectivity24–6 weeks20–40 points liveReal-time dashboard + utilizationPhase 1 acceptanceSTAGE 03Reporting & Traceability36–10 weeksWork orders · Reporting · QCFull-chain traceability of people, machines, materials, methodsPhase 2 acceptanceSTAGE 04Rollout to All Lines48–16 weeksReplicate to remaining linesUnified asset registry and metric definitionsPhase 3 acceptanceSTAGE 05Optimization & Expansion5OngoingAPS · Energy · Predictive maintenanceQuantifiable OEE improvementAnnual O&M
Five steps, each with its own acceptance criteria and its own value. Projects that skip step 2 and go straight to plant-wide MES usually end up eight months later with nothing but a dashboard nobody looks at.
Three red lines for judging whether a plan can be delivered: phase one lasts no more than 12 weeks; phase one must go live independently and deliver value; phase one must have one person accountable for the outcome (not a committee). If any one is missing, cut the scope down again.

IoT Application Scenarios Across Six Industries

The same platform becomes a different set of data points and rules in a different industry. The table below lists scenarios we have delivered or assessed, spelling out what gets collected and where the value lies.

IndustryWhat is collectedTypical applicationsWhere the value lands
ManufacturingPLCs, robots, special-purpose machines, energy metersWork order execution, utilization, quality disposition, predictive maintenanceLess downtime, higher first-pass yield, shorter changeovers
Logistics & WarehousingVehicle terminals, access control, temperature and humidity, weighbridgesReal-time fleet dispatching, route optimization, in-transit temperature control, dock queuingLess empty running, better on-time rates, less cargo damage
Agriculture & FoodWeather stations, soil sensors, greenhouse actuators, sorting linesClimate monitoring, greenhouse automation, batch traceability, sorting qualityLower spoilage, traceability, pricing power
Energy & EnvironmentElectricity and water meters, gas sensors, emission monitoringSub-metering, peak and valley optimization, exceedance alerts, carbon accountingLower energy costs, manageable compliance risk
Healthcare & LaboratoriesCold-chain probes, reagent inventory, equipment operating statusCold-chain break alerts, reagent expiry management, equipment utilizationPassing compliance audits, less consumable waste
Retail & StoresFootfall counters, POS, refrigeration cases, lighting and HVACFootfall analytics, merchandising optimization, remote equipment controlHigher sales per square meter, lower maintenance headcount

CASE STUDY

Related Case Studies

From a sorting line to a traceability system to full shop-floor digitalization — three deliveries at different levels of granularity.

Cloud-factory mobile app for a precision-electronics plant
Smart Factory IoTIntelligent manufacturing, precision electronics

Cloud-factory mobile app for a precision-electronics plant

A smart-factory mobile app (Android / iOS plus a web management backend) for a precision-electronics...

View details →
Clinical laboratory · reagent supply-chain management platform
Smart Factory IoTIntelligent laboratory, reagent management platform

Clinical laboratory · reagent supply-chain management platform

A reagent supply-chain management system for a hospital clinical laboratory, extended from the deliv...

View details →
Fruit & vegetable processing · intelligent sorting system
Smart Factory IoTSmart agriculture, smart manufacturing

Fruit & vegetable processing · intelligent sorting system

A grading and sorting line for a food-processing plant: hyperspectral imaging, multi-sensor fusion, ...

View details →

FAQ

Frequently Asked Questions

What if employees are unwilling to use it after it goes live?

This is the risk that is most often underestimated. Our approach: The workstation interface is designed with the participation of actual users, with on-site or remote supervision during the first two weeks of launch, and individual training by role with screen recordings kept. Link the actions that "must enter the system" to assessments (this requires internal promotion from you). Difficulty in using the system and employee resistance are often two sides of the same coin.

How much does the post-maintenance cost?

Defect repairs are free of charge during the warranty period (usually 12 months). After the warranty period, optional annual service packages or pay-per-use services are available for operation and maintenance. The billing items for both methods are clearly listed in the contract, and there is no vague "annual service fee" charged.

We have our own IT team. Can we work together with you?

Yes, it's possible, and it's usually smoother. The typical division of labor is: you are responsible for servers, accounts, networks, and internal process decisions, while we are responsible for application code, interfaces, and deployment. Upon delivery, we will provide interface documentation and deployment manuals, and you can take over maintenance independently afterwards. We will then shift to providing support on a per-use basis.

Have a fleet of machines to connect?Start with a feasibility assessment

Equipment list, protocols, collectible points, acquisition cycles — one page, clearly laid out, before we talk budget.

0.402776s