Home›AI›Industry AI›Digital Twin

XcellTwin|Digital Twin as a Service

A live copy of your plant, building or data centre that you can question and test.

XcellTwin connects the sensors, PLCs and meters you already have to a virtual model that stays in sync. See every asset's state, catch faults before they stop a line, and try a change on the twin before you try it on the floor.

OPC UA · MQTT · Modbus
Predictive maintenance
OEE & energy optimisation
What-if simulation
Twin graph & 3D views
Run 24×7 by our NOC
XcellTwin · Line 1 mirrorsample data
01Physical line and its twin, in step
9Twin types
4Maturity levels
3OT protocols native
24×7NOC & SOC
Tier IVMumbai hosting
ISO 27001 · ISO 20000-1 Microsoft Gold Partner Tier IV data centre, Mumbai 24×7 NOC & SOC MeitY empanelled Since 1999
What it is

A model that changes when the real thing changes. That is the twin.

A digital twin is a virtual model of a physical asset, process or place that is kept in sync with it by live data from sensors and control systems. A 3D drawing shows what a pump looks like. A twin shows what the pump is doing right now, how it got there, and what it is likely to do next.

  • Synced, not static. Values arrive from PLCs, meters and IoT sensors every few seconds, so the model is never a stale snapshot.
  • Connected. The twin knows the motor drives the conveyor that feeds the packer, so one fault can be traced to the assets it affects.
  • Testable. You can change a set-point, a speed or a schedule on the twin and see the effect before anyone touches the plant.
How a twin differs
CAD or BIM modelGeometry and specs. Accurate on the day it was drawn.
SCADA screenLive values for operators, tag by tag, with little context or history.
Offline simulationAnswers what-if questions, but with assumed inputs, not today's data.
Digital twinGeometry, live values, relationships and history in one model, with simulation run on real data.
Physical ↔ twin explorer

Click a machine. Read its twin.

A sample packaging line with five machines. The top row is the shop floor; the bottom row is the twin. Select a machine to see its live tags, then inject a fault to watch the twin raise it first. All values are simulated.

Line 1 · sample plant, Punein sync · lag 0.8 s
Physical assetTwin modelData in flightAlert
Twin types

Nine kinds of twin. Start with the one that costs you most.

Each tile shows what the twin mirrors, where its data comes from and a sensible first use case.

Asset

Asset & equipment twin

One motor, pump, compressor, CNC or transformer, with its health and history.

Data
Vibration, temperature, current, run hours
First win
Predictive maintenance on critical assets
Process

Production line & factory twin

Machines, buffers and people as one flow, so bottlenecks and losses show up.

Data
PLC states, counts, MES orders, quality checks
First win
OEE by line and downtime reasons
Place

Building & campus twin

Floors, zones, HVAC, lighting and meters in a single model of the space.

Data
BMS (BACnet, Modbus), energy meters, occupancy
First win
HVAC scheduling and energy per square metre
Place

Data-centre twin

Power chain, cooling, airflow and rack capacity, so you place load where it will run cool.

Data
PDU and UPS feeds, inlet sensors, CRAH units, CFD runs
First win
Hot-spot detection and capacity planning

Built by the team that runs XcellHost's Tier IV data centre in Mumbai.

Flow

Supply-chain & logistics twin

Plants, warehouses, lanes and stock as a network you can stress-test.

Data
ERP, WMS, TMS, GPS and cold-chain loggers
First win
What-if on a lost supplier or a closed lane
Flow

Network twin

Routers, links and policies modelled so a change can be checked before it is pushed.

Data
Configs, topology, SNMP and flow telemetry
First win
Change validation and failure what-ifs
Infra

Utility & infrastructure twin

Water networks, substations and feeders, with pressure, flow and load in one view.

Data
SCADA, RTUs, smart meters, GIS layers
First win
Leak and loss detection, load forecasting
Infra

Smart city twin*

Traffic, streetlights, water and waste on one map, for command-centre teams.

Data
Command-centre feeds, GIS, sensors, cameras
First win
Streetlight and water-network energy use

* Scoped per project with the city's system integrator.

Operations

Healthcare & hospital ops twin

Beds, theatres, equipment and patient flow, so delays are seen before they build up.

Data
HIS/ADT events, RTLS tags, BMS, equipment logs
First win
Bed turnaround and asset tracking
Maturity ladder

Four levels. Each one answers a harder question.

Most twins start at level one and climb. Pick a level to see what it needs and what you get.

Architecture

Six layers from sensor to screen. Open standards at every joint.

Hover or tap a layer. Product names are shown for compatibility; items marked * are third-party products we work with, not XcellHost products.

Built for India

Plants, cities and rules we already work with.

Industry 4.0 to 5.0

Twins are the data backbone for smart factories, and the people-centred Industry 5.0 work of cobots and operator assistance. See our Industry 5.0 programme and Manufacturing Cloud.

Smart Cities Mission

Launched on 25 June 2015 by the Ministry of Housing and Urban Affairs for 100 cities, the mission set up Integrated Command and Control Centres in every one of them. It formally closed on 31 March 2025; the command centres and their data remain, and are a natural base for city twins.

OT security · IEC 62443

We design twin connectivity to the IEC 62443 series: zones, conduits and least-privilege access between OT and IT. BIS has adopted parts of the series as identical Indian Standards, for example IS/IEC 62443-4-1:2018 for secure product development. See OT protection and OT assessment.

DPDP for people data

Hospital, building and city twins often touch personal data: badge swipes, CCTV, occupancy, wearables. That falls under the DPDP Act, 2023. We minimise it at the edge and map what remains with our DPDP compliance platform.

CERT-In reporting

CERT-In's directions of 28 April 2022 require reportable cyber incidents to be reported within 6 hours of noticing them. Our 24×7 SOC watches twin gateways and data paths and runs that clock with you.

Data stays in India

By default the twin platform runs in XcellHost's Tier IV data centre in Mumbai. You can also keep it at your plant edge or in partner data centres*, with rupee billing and GST invoices.

Use cases by industry

Same twin engine. Different first questions.

Predictive maintenanceVibration and current trends on motors, gearboxes and spindles flag wear days ahead.
OEE and loss treeAvailability, performance and quality per line, with downtime reasons from the PLC, not a clipboard.
Quality traceabilityLink each batch to the machine settings it ran with, so a defect points to a cause.
Operator trainingNew operators practise changeovers and fault drills on the twin, not on a live line.
Manufacturing Cloud → Industry 5.0 smart manufacturing →
Clean-room environmentPressure cascades, temperature and humidity per room, with excursions flagged and logged.
Batch what-ifTry a process parameter change on the twin before raising a change control.
Utility healthHVAC, purified water and compressed air monitored as the systems that keep batches in spec.
Audit-ready recordsTime-stamped tag history kept with access logs, to support your data-integrity controls.
Pharmaceutical Cloud →
Thermal hot spotsRack inlet heatmaps and airflow models show where cold air is short before an alarm fires.
Power chainUtility to UPS to PDU to rack, so you can see what a breaker trip would take down.
Capacity planningSpace, power and cooling headroom per row, before a new GPU cluster lands.
CFD on GPUsAirflow simulation runs on our GPU servers; GPU nodes are priced on quote.
XcellHost data centres → Data-centre build services →
HVAC optimisationSchedules and set-points tuned floor by floor against occupancy and outside temperature.
Energy and carbonMetered energy per tenant and floor, ready for ESG and BRSR evidence.
Facility maintenanceChillers, pumps and lifts on condition-based work orders instead of fixed calendars.
Construction handoverBIM models become the twin on day one, so the as-built record stays useful.
Construction Cloud → Hospitality Cloud →
Network what-ifClose a lane, lose a supplier or double a SKU's demand on the twin and read the service impact.
Warehouse flowDock, pick and pack modelled to find the congestion point before peak season.
Cold chainReefer and store temperatures streamed and checked against product limits.
ETA confidenceLive GPS and dwell times turned into arrival forecasts customers can trust.
Logistics Cloud → Retail Cloud →
Non-revenue waterPressure and flow by district metered area, to narrow down leaks.
Feeder load forecastSmart-meter and weather data forecast load on each feeder and transformer.
OT-safe remote opsRead-mostly twins let teams see remote sites without opening control paths.
City command centresTwin layers for streetlights, water and traffic on top of command-centre data.
IoT Cloud → Government Cloud →
Bed and theatre flowAdmission, transfer and discharge events show where patients wait and why.
Equipment trackingPumps, ventilators and wheelchairs located in real time, so staff stop searching.
Medical gas and powerOxygen, vacuum and backup power monitored as critical systems.
Privacy by designPatient identity stays in the hospital system; the twin works on events and locations.
Healthcare Cloud →
Delivery roadmap

Prove it on one line. Then scale what worked.

  1. 01

    Discover

    2–3 weeks*

    Workshop on site or online. We list assets, tags and protocols, pick one use case with a rupee value and agree the success measure.

    You get: use-case brief, tag map, pilot plan
  2. 02

    Pilot

    8–12 weeks*

    One line, one building or one data hall. Edge gateway installed, tags flowing, twin modelled, first alerts and dashboard live.

    You get: a working twin and a measured result
  3. 03

    Scale

    by phase*

    Copy the model to more lines and sites. Add 3D views, simulation and links to ERP, MES or CMMS.

    You get: a template you can repeat
  4. 04

    Run

    24×7

    Our NOC watches data freshness, gateways and models; our SOC watches the OT and IT paths. Monthly reviews of what the twin found.

    You get: a twin that stays accurate
Plans

Start with a workshop. Pay monthly once it runs.

All prices are new and indicative. GPU nodes for simulation and rendering are priced on quote.

Discovery workshop

Find the use case worth twinning first.

₹49,999*one-time
Indicative · excl. 18% GST
  • 2-day workshop with your plant and IT teams
  • Asset, tag and protocol inventory
  • Use case ranked by rupee value
  • Pilot scope and success measure
Most start here
Pilot twin

One line, building or data hall, live in weeks.

₹4,49,999*one-time
Indicative · excl. 18% GST
  • Edge gateway set-up · up to 500 tags
  • Twin model of one area
  • Anomaly alerts on critical assets
  • Dashboard and result review
Managed twin

Hosted, monitored and kept accurate.

₹24,999*/month
Indicative · excl. 18% GST · from 500 tags
  • 500 tags₹24,999*
  • 2,000 tags₹59,999*
  • 10,000 tags₹1,49,999*
  • Hosting in Mumbai Tier IV
  • Business-hours ops · 24×7 add-on
Enterprise

Many sites, 3D and simulation at scale.

On quote
Sized to sites, tags and models
  • Multi-site twin graph
  • OpenUSD 3D scenes and AR views
  • Custom ML and simulation models
  • Named twin architect
GPU nodes for simulation and renderingFor CFD, physics simulation, 3D scene rendering and model training. Sized to your workload on our GPU servers.
Price on quote

* Indicative — confirm with XcellHost. INR, excluding 18% GST. Durations are typical, not guaranteed. Sensors, PLC changes, site cabling, third-party software licences and travel outside Mumbai are quoted separately.

Configure

Build your twin. See the rupee total.

Fixed-price items go to the cart. GPU nodes and Enterprise are added to a quotation request instead.

1 · Managed twin size (tags)
4 · Where the twin runs
5 · Operations
6 · One-time and quoted items
Your estimate · indicative

XcellTwin · 500 tags

    Monthly₹0
    One-time₹0
    Subtotal · first invoice₹0
    GST (18%)₹0
    Total incl. GST₹0
    Tools

    Will a twin pay for itself? Run your own numbers.

    Both tools run in your browser and send nothing. Results are illustrations, not a forecast for your plant.

    Predictive-maintenance savings

    Illustrative calculator
    20
    24
    ₹50,000
    25%
    ₹30 L
    Downtime cost avoided₹0
    Maintenance saved (assumed 10%)₹0
    Year one: pilot + 12 months*₹0
    Net yearly benefit₹0

    —

    Twin readiness quiz

    6 questions · about one minute
    Compare

    Dashboards, 3D models, or a twin? What each one actually does.

    Question you need answeredSCADA dashboards3D model onlyXcellTwin
    What is each machine doing now?Yes, tag by tagNo live dataYes, in context of the line
    Which assets does this fault affect?~Operator knowledge~Shows layout onlyTraced through the twin graph
    Will this motor fail soon?~Fixed alarm limitsNoAnomaly and trend models
    What if we run 10% faster?Only by trying it live~Offline study, assumed inputsSimulated on today's data
    Where is the asset physically?~Schematic screensAccurate geometryGeometry linked to live tags
    Can IT, finance and plant teams share it?Inside the control room~Needs CAD or BIM toolsBrowser, role-based access
    Who keeps it running?~Your controls team~Drawn once, then agesXcellHost 24×7 NOC

    A twin does not replace SCADA. SCADA keeps controlling the plant; the twin reads from it, adds context and runs analysis away from the control loop.

    Frequently Asked Questions

    A virtual model of a physical asset, process or place that is kept in sync with it by live data. Unlike a drawing or a one-off simulation, it updates as the real thing changes, keeps history, and knows how assets relate to each other, so you can monitor, diagnose, forecast and test changes on it.

    Usually not for a pilot. Most plants and buildings already have PLCs, SCADA, BMS and energy meters with useful tags. The discovery workshop lists what you have. We add sensors only where a use case needs them, for example vibration sensors on critical motors, and quote them separately.

    No. By default the twin reads data through an edge gateway and never writes to PLCs or SCADA. Connections follow IEC 62443 zone and conduit principles, with one-way or read-only paths where your OT team requires them. Any write-back, such as set-point recommendations, stays with an operator to approve.

    OPC UA, MQTT and Modbus (TCP and RTU) at the edge, plus BACnet for buildings and SNMP for IT and data-centre gear. Upstream we connect to historians, MES, ERP and CMMS through their APIs or files. Uncommon drivers are checked in the discovery workshop.

    We can. Azure Digital Twins models twins in DTDL, a JSON-LD-based language, and supports DTDL v2 and v3. AWS describes IoT TwinMaker as a service that builds a knowledge graph of your assets with 3D scenes and a Grafana plug-in; we check AWS's service lifecycle notices before designing on it. NVIDIA Omniverse builds on OpenUSD for 3D scenes and simulation. If you already use a cloud platform we build on it; otherwise we use open-source components hosted by XcellHost.

    OpenUSD (Universal Scene Description) is an open-source framework, created by Pixar, for describing and combining 3D scenes. The Alliance for OpenUSD was formed in 2023 by Pixar, Adobe, Apple, Autodesk and NVIDIA to standardise it. Using it keeps your 3D twin portable between tools instead of locked into one vendor's format.

    A discovery workshop takes 2–3 weeks* end to end. A pilot on one line, building or data hall typically takes 8–12 weeks* from kick-off to a measured result, depending on site access and how quickly tags can be exposed. We give you a dated plan after discovery.

    By default in XcellHost's Tier IV data centre in Mumbai. You can keep the platform at your own plant edge, or in partner data centres* where that suits your network. Time-series and history can also land in a managed data lake for wider analytics.

    Not for monitoring, alerts or dashboards. GPUs help with CFD airflow models, physics simulation, rendering large 3D scenes and training bigger ML models. GPU nodes are always priced on quote, sized to the workload, on our GPU servers.

    If the twin processes digital personal data, yes. Machine and energy data usually is not personal data, but badge swipes, CCTV, occupancy by person and worker wearables can be. We aggregate or drop identifiers at the edge where possible and help you document the rest.

    Indicatively ₹49,999* for a discovery workshop, ₹4,49,999* for a pilot twin, and a managed twin from ₹24,999* a month for up to 500 tags. Enterprise programmes and GPU nodes are quoted. All prices exclude 18% GST. Use the configurator for a total.