AGV automatic charging system selection starts with a practical question: when and where does each vehicle stop, for how long, and how much energy can its battery actually receive before the next transport order? Charger nameplate power alone cannot answer it. A Thai factory must account for material handoffs, shifts, floor construction, electrical capacity, traffic, and fleet software together. This guide shows how to compare contact docking, contactless opportunity charging, and a hybrid arrangement using operating logs. It provides RFP questions and FAT/SAT acceptance tests. For broader deployment context, see our AGV/AMR introduction guide and AGV layout guide.
Define the stops and energy demand before choosing a charger
In a September 2026 article, OMRON Robotics distinguishes a dedicated docking trip from opportunity charging at recurring process stops. It cites conveyor transfers, material handoffs, machine loading, and staging as possible locations. The same manufacturer says conventional docks remain suitable for many sites. Treat this as a supplier’s explanation, not a guarantee that wireless charging improves every fleet. A stop must be frequent, long, and repeatable enough; the robot must park within the charger’s tolerance and the battery must accept the proposed power.
Collect a representative operating trace with vehicle and task IDs, load, route, arrival and departure times, reason for stopping, charging start and end, state of charge (SOC), and alarms. Include high-load days, shift changes, breaks, material shortages, line stops, and restarts. A daily average hides simultaneous charging requests. For a new fleet, record the assumed route and task schedule in the RFP and require a measured update after the pilot.
Score each candidate location on five questions: Is it already a process stop or a detour? How variable is dwell time? Can the vehicle align despite load, floor and map changes? Will a parked vehicle obstruct people, forklifts, maintenance or emergency access? Are power, cabling, cleaning and service access realistic? A frequently visited point can still deliver little energy if dwell is too short. A long wait may also be unusable if urgent tasks often interrupt it.
Close the energy balance in Wh
For a vehicle and a shift, call consumed energy E_use, energy accepted by the battery E_charge, and the change in stored energy ΔE. Conceptually, ΔE = E_charge − E_use. The design must survive the most demanding continuous work interval without crossing the battery maker’s minimum SOC. Displayed SOC is an estimate. Temperature, aging, load, speed and floor condition affect usable energy. Request usable capacity, recommended SOC window, charging limits, thermal derating, and BMS/charger communication requirements for the exact vehicle and battery.
Here is a hypothetical calculation, not a supplier measurement. If a vehicle pauses eight times per shift for six minutes and accepts a net 1.5 kW at the battery during each stop, the ideal added energy is 1.5 kW × 6/60 h × 8 = 1.2 kWh. Alignment delay, startup delay, taper at high SOC and communication interruptions reduce real energy. Ask vendors for measured battery-side Wh over your dwell-time and SOC ranges, not just rated charger output.
Compare contact, contactless and manual charging on one basis
| Criterion | Contact dock | Contactless opportunity charging | Manual plug or battery change |
|---|---|---|---|
| Best candidate | Predictable longer stop and dedicated space | Frequent process stops and compatible vehicle/floor equipment | Small fleet, short operating hours or phased retrofit |
| Verify on site | Contact alignment, contamination, wear, cleaning and recovery | Transmitter/receiver tolerance, floor works, foreign objects and nearby equipment | Operator time, errors, handling, storage and procedure |
| Fleet logic | Travel to dock, reservation and exit | Short sessions integrated with tasks | Human work planning and restart confirmation |
| Main cost/maintenance issue | Contact upkeep and dock queue | Compatibility, actual received power and construction | Labor, safety and missed manual steps |
MiR Charge 48V is an example of a manufacturer-specified automatic charging station. OMRON’s LD specifications list robot and dock conditions by model. These examples show that “automatic charging” is not one universal interface. Confirm the exact vehicle, battery, charger and software versions supported under the warranty. A connector that physically fits does not prove charging-profile, BMS, protective-shutdown or liability compatibility.
For a contact dock, test the worst credible parking offset, approach angle, floor wear, wheel state and load. OMRON’s LD data gives different repeatability conditions for a single robot and a fleet, with further differences by positioning method. Those values are model-specific; the lesson is to link charger tolerances to the actual navigation and alignment configuration.
For wireless charging, evaluate receiver alignment, floor-unit installation, cleaning, foreign-object management, heat, EMC, cabling and neighboring equipment. OMRON describes in-ground, raised-floor and on-ground arrangements. Wiferion’s manufacturer downloads provide product information, but obtain the current data sheet and installation limits for the proposed model. Removing exposed contacts does not remove maintenance of the floor unit, communications, cooling and positioning.
Charging queues depend on arrival times, not just fleet-to-dock ratio
Two chargers for ten AGVs do not automatically mean sufficient capacity. Breaks, shift changes and line restarts can synchronize arrivals. Detour distance, the space in front of a dock, whether charging can be interrupted, and when the fleet releases a vehicle all matter. Check reservation, task priority, low-SOC assignment and traffic controls in the deployed fleet-software version. OMRON’s FLOW Core page is one primary product reference, not proof that a particular installation includes every feature.
A simple first simulation uses 15- or 30-minute buckets: vehicles requesting a charger, occupied stations, departures and undelivered jobs. A queue that keeps growing calls for comparison of added stations, earlier charging, station relocation, task priority and process buffers. If transport demand itself exceeds fleet capacity, another charger cannot solve the underlying problem.
Size and place AGV charging stations in three stages
First build a dedicated-dock baseline, including travel and queue time, electrical capacity, floor works, maintenance access and minimum SOC under peak transport demand. Second model opportunity charging only where vehicles already stop. Apply measured dwell distributions, alignment success, real battery-side power and SOC-dependent limits. Third model a hybrid: process-stop top-ups plus longer sessions during breaks or low demand. Define how tasks and vehicles move when either charging path fails. More options can improve resilience but also add equipment and control complexity.
Compare the alternatives over the same period. Include chargers, installation, electrical and network works, controls integration, spares, service, battery replacement and lost production from interruptions in addition to vehicle cost. Our AGV cost guide provides the broader purchasing context. Calculate tariffs and contracted power from the specific Thai facility’s agreement and load profile. A universal price or payback claim would conceal site differences.
The layout drawing should show more than hardware. Mark approach and exit directions, vehicle/load envelopes, power cabinets, protection, cables, network equipment, pedestrian paths, forklift crossings, evacuation and maintenance access. For an in-ground installation, check cutting, waterproofing, cure time and floor transitions. For an on-ground platform, check trip and traffic hazards. Trace where a vehicle stops after a power or network failure.
Twelve answers the RFP should require
- Supported configuration: exact robot, battery chemistry, BMS, charger and software versions, territory and warranty boundary.
- Operating trace: fleet size, load/task distributions, shifts, peak definition and expansion scenario.
- Energy evidence: measured or conservative Wh per trip/hour, usable capacity, temperature and aging assumptions.
- Charging performance: battery-side power by SOC and temperature, startup delay and Wh delivered during short stops.
- Alignment: allowed position/angle error, approach, retry and safe behavior on misalignment.
- Station count: peak queue, minimum SOC, growth margin and a one-station-failed case.
- Fleet logic: reservations, priorities, low-SOC task allocation, interrupted charging and manual recovery.
- Electrical, floor and network: maximum concurrent demand, protection, earthing, drawings and offline behavior.
- Safety and environment: risk assessment, applicable standards, liquids, dust, temperature, cleaning and EMC.
- Maintenance: inspections, consumables, replacement time, diagnostics, spares and support scope.
- Acceptance: FAT/SAT conditions, instruments, logs, thresholds and retest responsibility.
- Handover: drawings, backups, bill of materials, versions, procedures and training in languages the local team uses.
Give every bidder the same task trace, load, temperature and battery-aging assumptions. Ask for inputs, calculation method, outputs and exclusions separately. Use mandatory gates for supported compatibility, available facility power, assessed safety functions and measurable logs. Then score compliant proposals on total cost, transport service, schedule, maintainability and expansion. A demonstration video cannot replace a test with representative workpieces and real stops.
Make the charging-queue model reproducible
Give suppliers time-stamped transport jobs with origin, destination, workpiece weight, priority and due time. Provide each vehicle’s initial SOC, usable capacity, speed, loaded consumption, idle consumption and charging limits. Add station locations and capacities, alignment/startup delays and planned outages. Keep supplier estimates separate from plant measurements. A coarse time step may hide a six-minute charging opportunity; a detailed model with wrong inputs is no better. The aim is to compare alternatives under common loads, not predict every future minute.
Run at least normal demand, maximum demand, one charger out of service, simultaneous shift return and a line-stop restart. Report delayed tasks, minimum SOC, median and long-tail charging wait, station occupancy and charging detours. Calibrate consumption and charging curves with repeated light- and heavy-load runs, short and long stops, and battery-side power measurements. Retest on a different day or shift from the calibration data. Vary demand, battery capacity, startup delay and dwell time to find where service first breaks down. That sensitivity tells the buyer whether the next investment belongs in electrical capacity, station placement, fleet software or operations.
Assign responsibility by interface
A failed charge may originate in the supply, floor transmitter/contact, receiver, BMS permission, fleet reservation or network. Draw the sequence from the facility power connection to task return. For each interface specify document version, signals, timeout, alarm code, log owner, first responder and final resolution owner. If multiple vendors participate, assign one integration-test lead. The plant’s preparations—distribution-board space, earthing, floor strength, Wi-Fi design, security review and production downtime for installation—belong in the same responsibility matrix.
Handover should include one-line electrical diagrams, installation drawings, settings and backups, parts lists, alarm codes, maintenance procedures and training records. Tie performance acceptance to agreed transport scenarios and logs. Define what falls outside the guarantee and who pays for retesting after external-system changes. This prevents a charger vendor and a fleet integrator from each treating a cross-system failure as somebody else’s issue.
Read safety standards within their actual scope
ISO 3691-4:2023 is the published standard for safety requirements and verification of driverless industrial trucks and their systems. Its ISO page explicitly says power-source requirements are outside its scope. It is therefore not a blanket certification of a charger or battery. A newer 2026 DIS revision is a draft, not a replacement published edition. IEC 62485-3:2014 covers safety for traction batteries and their installations, including AGV applications. It does not substitute for every charger approval. Check Thai electrical and workplace requirements, battery-maker instructions, vehicle conditions and the site’s risk assessment as applicable.
Consider electric shock, short circuits, overheating and battery faults alongside wrong docking, blocked traffic, unexpected departure, maintenance under power, liquid ingress and recovery after an outage. FAT checks controls and protection; SAT checks the real floor, traffic and escape routes. Define the responsible assessor and evidence for each hazard rather than listing standards alone.
FAT/SAT: prove an operating service, not just a charging lamp
| Test group | FAT | SAT | Evidence |
|---|---|---|---|
| Compatibility | Exact vehicle/BMS/charger/version | Every installed vehicle and settings backup | Model, version, communications log |
| Charging | Power and startup delay by SOC/temperature | Battery-side Wh in actual stops | Meter, SOC and time log |
| Alignment | Center and edge of allowed range | Real floor, load and traffic | Stop position and failures |
| Peak operation | Reservation and queue with test fleet | Undelivered jobs and minimum SOC under representative load | Jobs, queue, SOC trace |
| Fault recovery | Power, network, contact/foreign-object faults | Outage, disconnection and restart | Alarm, shutdown and recovery log |
| Maintenance | Inspect, clean and replace | Local team repeats procedures | Procedure, duration and training record |
Set numeric thresholds for the selected equipment and process. For example, agree limits for undelivered jobs during a defined two-hour peak, battery maker’s minimum SOC, and delivered Wh per planned stop. Repeat borderline tests and retain the lower results as well as averages. A pass at one site is not automatically a pass at another.
Design failure tests early. What happens if a low-SOC robot receives a high-priority task, its reserved charger fails, an urgent job arrives mid-charge, the network drops while power returns, or maintenance isolates a station? Verify safe stopping, alternate charging, task reassignment and manual recovery. Log robot ID, station ID, request/arrival/alignment/charge start and end, battery-side Wh, SOC, temperature, cutoff reason and time back to work. Synchronize clocks and event IDs between charger and fleet systems. If energy totals disagree, check whether a meter reports grid-side or battery-side energy.
Classify a failed test as equipment, vehicle, construction, communication, fleet logic or test-condition failure. Record the owner, correction, retest scope, deadline and temporary operating restriction. Cleaning a contact once is not a permanent fix until the contamination source and maintenance interval are understood. A wireless positioning failure may require changes to navigation or mapping, not only the floor unit.
Operating KPIs and triggers for redesign
A busy charger can coexist with missed transport. Review each vehicle’s minimum SOC, request-to-charge wait, Wh per stop, charging detour, station occupancy, failed charging, low-SOC job delays and maintenance downtime weekly. Review battery capacity, contact/transmitter condition and costs monthly. Recalculate the plan after fleet expansion, longer shifts, battery replacement, software updates, floor works or a moved handoff point. “Expandable” only has meaning if the electrical board, network, station space and reservation logic can expand too.
Write procedures for charger faults, night/weekend support, manual fallback, isolation, spares and outage recovery. Derive KPI targets from the plant’s transport service level. Count jobs delivered late because of charging separately from delays caused by material shortages or other process problems.
FAQ: AGV automatic charging system selection
How many AGV charging stations do we need?
There is no reliable fixed ratio. Use the actual energy demand, operating hours, accepted charge power, stop distribution, synchronized requests and dock detours. Test peak queue and minimum SOC, including one failed station, against the same trace for all bidders.
Is contact or contactless charging better?
Contact docking may fit predictable longer stops and a dedicated area. Contactless opportunity charging may fit recurring process pauses where the vehicle, floor and power system are compatible. A hybrid may be appropriate. Compare delivered battery-side Wh, queue, construction and maintenance rather than a technology label.
Can fleet software remove charging queues?
Reservations, priorities and low-SOC assignment can reduce avoidable waiting. They cannot create capacity when simultaneous demand exceeds the available stations and battery charging limits. Compare controls with additional stations, relocation and operating changes.
What matters most in AGV charging FAT/SAT?
Replay the representative peak and measure minimum SOC, queue, delayed transport and battery-side Wh together. Then separately test misalignment, power/network loss and recovery from a failed station under the applicable safety assessment.
Does ISO 3691-4:2023 certify charger safety?
No. It addresses driverless industrial trucks and systems and excludes power-source requirements. Evaluate the battery and charging installation under appropriately scoped standards such as IEC 62485-3:2014, local requirements and manufacturer instructions. Do not present a 2026 DIS draft as a published standard.
Conclusion: accept the transport service, not only the charger
Use real stops and battery-side Wh to close the energy balance. Model peak queues and minimum SOC, compare contact, contactless and hybrid options under identical assumptions, and test faults in FAT/SAT. Procurement should demand evidence that the fleet can keep the factory’s required material flow running.
If you are assessing existing AGV charging queues or the station count for a new Thai AMR fleet, we can help define the comparison from operating logs and a layout, even at an early planning stage. Contact TOMAS TECH.
