Havish M Consulting
Customer Case Study · Supply Chain · Plant Logistics
Consumer Durables Manufacturer × Havish M Consulting

Every truck, dock and bay, live on one screen

A 3D logistics digital twin of a room-AC plant. It follows every supplier truck from the gate to the dock, every packed lot from the final line to the dispatch yard, and lets the plant test peak-season what-ifs before a single truck queues.

ClientConsumer Durables Manufacturer FunctionSupply Chain & Plant Logistics ThemeLogistics Digital Twin
Inbound truck turnaround, gate to gate
−0%
▼ 3 h 05 min → 1 h 55 min
Dock use while trucks are waiting
0%
▲ from 64% in the last peak season
Truck detention charges
−0%
▼ peak season vs peak season
Time to locate any truck on site
<0min
▼ from about 20 min of phone calls

Figures indicative; validated metrics available on request.

The Challenge

Peak season meant queues at the gate and idle docks inside

In AC season (January to April) inbound steel, copper tubes, compressors and electronics roughly double, while the plant ships thousands of AC units a day to brand customers. All of it moves through one inward gate, six docks and one finished-goods yard.

Queues next to empty docks

Supplier trucks waited outside for hours while two of six docks sat idle. Security could see the queue, stores could see the docks, nobody saw both.

Four registers, no single truth

Gate pass, weighbridge slip, GRN and dispatch register each lived in a different book or WhatsApp group. Finding one truck took a round of phone calls.

An overflowing FG yard

Loaded trucks for brand customers waited for documents, seal checks and transporters. The yard filled up and detention charges climbed every peak season.

Lost at shift change

Three shifts, verbal handovers. A dock blocked for GRN at 13:50 was still blocked at 15:00, and the next shift found out when a truck could not dock.

HMC Approach

From gate registers to a twin the whole plant watches

Built with the plant's security, stores, PPC and dispatch teams, starting from their own gate registers, weighbridge slips and GRN timestamps.

Select a step to see what happened.

Discover · Walk the truck

We followed trucks from the highway to the dock and from the line to Gate 2.

  • Time-stamped about 600 truck visits from gate registers, weighbridge slips and GRNs.
  • Split every visit into steps and found where the waiting sat: parking, GRN holds and documents.
  • Mapped the site: gates, weighbridges, 12 parking bays, 6 docks, 24 dispatch bays.

Deliverable: Truck-visit anatomy and a site layout model.

Design · Model the plant

Zones, docks, bays and the rules that move trucks between them.

  • Dock allocation rule: next free dock, yard trucks called in arrival order.
  • Holds and pauses: GRN blocks, forklift holds, shift changeovers.
  • KPI tree agreed with the plant head: turnaround, dock use, yard occupancy, detention.

Deliverable: Twin specification and KPI tree.

Build · A twin you can watch

A 3D model of the whole plant with an event engine underneath.

  • Three.js twin: gates, weighbridges, docks, shops, final line, QA and FG yard.
  • Simulated trucks follow real road paths, queue, reverse into docks and wait for GRN.
  • Built with Claude as the coding partner; dock what-if simulator calibrated on the visit data.

Deliverable: Plant logistics twin (single HTML) and what-if simulator.

Adopt · Run the day from the twin

The twin went on screens at security, stores and the dispatch office.

  • 08:00 dock plan reviewed on the twin each morning.
  • Blocked-dock alerts to stores; yard trucks called by PA and WhatsApp.
  • Auto-drafted shift brief at every handover.

Deliverable: Control-room routine and shift-brief template.

Scale · Live feeds and more plants

From simulated flows to live data, plant by plant.

  • Gate-pass, RFID and weighbridge feeds replace manual entries.
  • Supplier dock-slot booking piloted with copper and compressor vendors.
  • Same twin template for other plants.

Deliverable: Data-feed roadmap and multi-plant template.

The Solution

The plant logistics twin

The twin itself is embedded below and running. Take the guided tour, switch to a peak-season scenario, then use the simulator to decide how many docks the season really needs.

01 · Live twin

The whole plant, gate to gate

Drag to pan, scroll to zoom, click any truck or building. The twin runs one plant minute per second; use the controls inside it to speed up, switch to night or show truck paths.

Loading the plant twin…The 3D model builds when this section comes into view. It needs WebGL and an internet connection for the 3D library.
Normal day6 docks · supplier truck every ~20 plant minutes · takt 30 min

Guided tour

Pick a stop to fly the camera there, or play the whole tour.

Plant clock
—
waiting for the twin
Trucks on site
—
—
Waiting (gate + yard)
—
trucks
Dock utilisation
—
AC units packed
—
—
FG yard
—
bays occupied
Live feed · queue and dock use
Telemetry streams from the twin once it is running.

The twin runs a simulated day. Supplier descriptions, truck IDs, destinations and volumes are illustrative; in production, events come from gate, weighbridge and ERP feeds.

02 · Shift brief

The handover writes itself

At every shift change the twin's numbers and event log are turned into a one-page brief with the actions for the next shift. Press the button to draft one from the twin running above.

What the brief reads

Live KPIs: trucks, queue, docks, FG yardBlocked docks and GRN holdsLine output in AC units and truckloadsThe last events in the twin's log

Rules it applies

Queue above 3, dock blocked, yard above 70% or turnaround above 2 h each trigger a specific action for the next shift.

Rule-based sample built from the twin's live numbers. In production Claude drafts it from the full event log.

Shift handover briefDraft

Press “Draft shift brief” to generate a handover from the twin's current state.

03 · Dock what-if

How many docks does peak season need?

Simulate a full day of inbound trucks, twice: the old way (paper registers, peaky arrivals, manual dock calls) and with the twin (automatic allocation, slot-booked arrivals). Every block on the board below is one truck at a dock.

Plant inputs

Steel, copper, compressors, motors & PCBs, R32
Paperwork or quality hold before the dock is released

Turnaround by number of docks

BeforeWith the twin
Dock board · with the twin
Trucks waiting for a dock, through the day
BeforeWith the twinArrivals per hour (shown day)

Indicative discrete-event simulation. "Before" assumes peaky arrivals, 15 min at the gate and manual dock calls; "with the twin" assumes slot-booked arrivals, pre-registered trucks and automatic allocation. The live version is calibrated on the plant's own visit data. Detention is estimated at ₹250 per truck-hour beyond 3 hours, 26 working days a month.

Impact Dashboard

Less waiting, not faster unloading

Unloading time hardly changed. What disappeared was the waiting around it: in the parking yard, on GRN holds and for documents in the FG yard.

Anatomy of a truck visit

Average minutes per step, peak season

WaitingValue-addingProcess steps

When the gate queue builds

Average trucks waiting at gate and yard, by hour and weekday

Figures indicative; validated metrics available on request.

Before vs After

How a truck moves through the plant

DimensionBeforeAfter
Truck visibilityGate register, weighbridge slips and phone callsEvery truck live on the twin with cargo, dock and time on site
Dock allocationSupervisor walks the yard and calls the next truckNext free dock assigned automatically; yard truck called by PA and WhatsApp
GRN holdsFound when the next truck cannot dockDock turns red on the twin and stores are alerted in a minute
Peak planningExtra docks and labour decided on gut feelSeason's arrivals run through the simulator before money is spent
FG dispatchLoaded trucks wait for documents; transporters chased by phoneYard board by truck size and destination; documents raised at QA
Shift handoverVerbal, and lost between shiftsAuto-drafted brief from the twin's numbers and event log
Tools & Tech Used

Built light, built fast

Three.js 3D plant twin Claude coding partner, event logic & shift briefs Python (SimPy) dock & yard simulation, calibration D3.js dashboards & simulator Gate & weighbridge logs 12 months of truck visits ERP GRN timestamps dock dwell & holds WhatsApp / PA alerts dock calls & blocked docks
Key Takeaways

What transfers to any plant gate

01

Make the waiting visible

Most of a truck's visit was invisible waiting, not slow unloading. Putting every truck on one screen cut it before any new dock was built.

02

Simulate the season, then spend

Running peak arrivals through the simulator showed when a seventh dock pays back and when slot booking alone is enough.

03

Let the twin write the handover

An auto-drafted shift brief kept context across three shifts and gave the plant head one page every morning.