De combien de couple un AGV a-t-il besoin? Une réponse basée sur la charge utile avec des normes d'ingénierie

De combien de couple un AGV a-t-il besoin? Une réponse basée sur la charge utile avec des normes d'ingénierie

Engineers designing Automated Guided Vehicles (AGV) et robots mobiles autonomes (RAM) routinely ask one question early in the project: how much torque does the drive motor actually need? The answer is never a single number. It depends on vehicle mass, acceleration target, slope angle, diamètre de roue, floor friction, number of driven wheels, gearbox ratio, and the thermal duty cycle the motor must sustain. This guide breaks the question down by payload class, provides force-model formulas you can apply immediately, and grounds the recommendations in international motor standards (CEI 60034-1, PAS DE MG 1) and manufacturer technical data from Maxon, Faulhaber, et Yaskawa.

Whether you are building a 50 kg indoor AMR or a 60-ton heavy-load transfer cart, the methodology below will get you to a defensible motor torque specification before you issue an RFQ.

Why Torque Sizing Determines AGV Success or Failure

Undersized torque causes motor stall under load, overheating during sustained operation, and failure to climb ramps. Oversized torque wastes battery capacity, increases wheel slip on smooth floors, and raises BOM cost without performance gain. In production environments, AGV reliability is rarely limited by control software — it is limited by the mechanical interaction between motor torque, wheel traction, and floor conditions [1].

The torque requirement is not a single value but a profile across operating conditions. An AGV that moves smoothly on flat epoxy may stall on a 3% ramp. A motor that handles straight-line cruising may fail during in-place rotation, which typically demands two to five times the straight-line torque in differential-drive configurations [5].

The Quick Answer: Torque by Payload Class

If you need a ballpark before diving into formulas, the table below maps typical AGV payload classes to per-wheel continuous torque, recommended motor type, voltage platform, and gearbox ratio. These values assume 2 driven wheels, polyurethane tires on smooth concrete (Crr = 0.02), 0.5 m/s² acceleration, sol plat, and a safety factor of 1.5. They are starting points, not final specifications — always validate with the force model in the next section.

The torque values in Table 2 are per driven wheel. For a 2-wheel differential drive, the total tractive torque is double the per-wheel value. For a 4-wheel-drive configuration, divide the per-wheel value by 2 (assuming equal load distribution). Always recalculate using the formulas below for your specific operating conditions.

Force Model: Four Resistance Components

AGV drive torque is determined by the total resistance force the vehicle must overcome. The force model decomposes this into four components, each corresponding to a physical resistance source. This decomposition follows the methodology described in the MDPI Engineering Proceedings paper on motor parametric calculations for robot locomotion [9] and is consistent with the sizing approaches used by Oriental Motor [3] and iNetic Motion [4].

1. Résistance au roulement (F_rouler)

Rolling resistance is the force required to deform the tire and floor at the contact patch. It depends on the tire material, floor surface, and normal load. Polyurethane tires on smooth epoxy have the lowest rolling resistance; rubber tires on rough concrete have the highest.

Formule: F_rouler = m × g × C_rr × parce que(je)

Où m is gross vehicle mass (kg), g = 9.81 m/s², C_rr is the rolling resistance coefficient, et je is the slope angle. Sur terrain plat, cos(0°) = 1, so the term simplifies to m × g × C_rr.

2. Acceleration Force (F_acc)

Newton’s second law: the force required to accelerate the vehicle’s mass from rest to target speed. This is typically the largest force component during startup.

Formule: F_acc = m × une

Où un is the target acceleration (m/s²). AGV acceleration is usually 0.3–0.8 m/s² for stability. AMRs may reach 1.0 m/s². Emergency deceleration can require 1.5–2.0 m/s², which produces the highest force transient.

3. Résistance de qualité (F_grade)

When the AGV operates on a ramp — loading dock approaches, floor transitions, or inclined transport paths — gravity adds a component parallel to the slope. This force can be substantial even on modest grades.

Formule: F_grade = m × g × péché(je)

UN 1,200 kg AGV sur un 3% ramp must overcome 353 N of grade resistance alone — nearly equal to its rolling resistance on flat ground. If your AGV encounters ramps, grade resistance often becomes the dominant force component.

4. Résistance au virage (F_tourner) — Differential Drive Only

In differential-drive AGVs (deux roues motrices, plusieurs roulettes), in-place rotation generates the highest torque demand. Caster wheels must pivot, creating significant scrub resistance. An engineering approximation from field data [5]:

Formule: F_rotation = (2 ×F_rouler × √(W² + L²)) / O

Où O is wheel track width and L is vehicle length. En pratique, rotation torque is 2–5× the straight-line torque, and it usually determines the peak torque rating of the motor.

Core Torque Formulas and Variables

Combining the four force components, the total driving force and wheel torque are calculated as follows. This formulation is consistent with the method described in the MDPI Engineering Proceedings paper [9] and the AGV Drive Wheel sizing guide [2].

Total Driving Force

F_total =F_rouler + F_acc + F_grade

(Turning resistance is evaluated separately as a peak condition, not added to the continuous force.)

Per-Wheel Torque

T_roue = (F_total ×r_roue) / n_conduire

Où r_roue is the chargé rayon de roue (not the nominal radius — a polyurethane tire compresses 2–5 mm under load), et n_conduire is the number of driven wheels sharing traction.

Motor-Side Continuous Torque

T_moteur = (T_roue × SF) / (je × n_g)

Where SF is the safety factor (1.25–1.5 for indoor, 2.0 for heavy industrial, 2.5 for safety-critical), je is the gearbox reduction ratio, et le_g est l'efficacité de la boîte de vitesses.

Pour les applications AGV, two-stage planetary gearboxes are the most common choice because they offer the best balance of efficiency, densité de couple, and backlash. For an in-depth comparison of gearbox types, voir notre spur gear motor vs. moteur à engrenage planétaire analyse. Worm gearboxes are generally avoided in AGVs due to their low efficiency (which wastes battery capacity) and self-locking behavior (which prevents coasting and regenerative braking).

Power Check

P_roue =F_total ×v
P_moteur =P_roue / (le_g ×n_moteur)

Où v is target travel speed (MS), and η_moteur is the motor efficiency (0.85–0.92 for BLDC at rated load). This power figure should include a 30–50% margin for surge, braking hold, and ramp startup.

Exemples travaillés: 150 kg AMR and 1,200 kg AGV

Exemple 1: 150 kg RAM (Goods-to-Person Robot)

Résultat: UN 150 kg AMR requires approximately 0.36 N·m continuous motor torque per drive wheel with a 20:1 réducteur planétaire. UN 24Dans le moteur BLDC in the 200–300 W range with a 42–57 mm frame size is appropriate. Le complete AGV motor selection guide provides additional payload classes and motor model recommendations.

Exemple 2: 1,200 kg AGV (Assembly Line Transport)

Résultat: Le 1,200 kg AGV requires approximately 3.04 N·m continuous motor torque per wheel with a 30:1 boîte de vitesses. A 48V BLDC motor in the 1–2 kW range (86–115 mm frame) is appropriate. Note that the grade resistance (353 N) contributes 30% of the total force — if the AGV operates only on flat ground, the required torque drops to 2.2 N·m and the power to 1,540 O. This highlights why you must size for the worst-case operating point, not the average.

For a deeper treatment of torque calculation methodology, including differential-drive turning torque and inertia matching, voir notre Guide de calcul du couple du moteur AGV with full force models and standard references.

Payload-Based Motor Selection Matrix

The table below synthesizes the calculations from the worked examples and extends them across the full payload range. It assumes 2-wheel differential drive, polyurethane tires on smooth concrete, 0.5 m/s² acceleration, and includes both flat-ground and 3% grade scenarios.

For custom motor specifications outside these standard payload classes, GreenSky Power offers custom electric motor design with frame sizes from 22 mm to 120 millimètre, voltage options from 12V to 72V DC, and integrated gearbox solutions.

Validation thermique: CEI 60034-1 Cycles de service

Torque alone does not guarantee motor survival. The motor must sustain the required torque within its thermal limits over the actual duty cycle. CEI 60034-1:2022 (Edition 15, published March 2026) définit dix classifications de cycles de service, of which five are most relevant to AGV applications [7].

Most AGV applications fall under S3 or S4 duty. The key distinction: if your AGV starts and stops frequently (typical cycle: 10 seconds moving, 20 seconds loading), the motor winding does not fully cool between cycles, and the continuous torque rating must cover the RMS torque over the full cycle, not just the peak.

RMS Torque Calculation

For intermittent duty, calculate the RMS torque over one complete cycle:

T_RMS = √[(T₁²×t₁ + T₂²×t₂ + … + T_n²×t_n) / (t₁ + t₂ + … + t_n)]

The motor’s rated continuous torque must exceed T_RMS at the operating ambient temperature. If T_RMS exceeds the rated torque, the motor will overheat — even if the peak torque is well within the motor’s capability.

Thermal Derating by Ambient Temperature

Motor torque ratings in datasheets are specified at 25°C ambient (per Maxon standard specification 100/101) [11]. À des températures ambiantes plus élevées, the permissible continuous torque decreases.

If your AGV operates in a 40°C ambient (common in un-air-conditioned warehouses), you must derate the motor by 10–15%. Un moteur conçu pour 5 N·m continuous at 25°C delivers only 4.25–4.50 N·m at 40°C. For high-temperature environments, specify Class F (155°C) ou Classe H (180°C) isolation, which allows 100% rated current up to 50°C ambient [12].

PAS DE MG 1 Torque Classifications for AGV Motors

PAS DE MG 1-2021 classifies motors into four design types based on torque characteristics and starting-load inertia. While NEMA standards are primarily used for AC induction motors, the torque classification framework is useful for understanding motor behavior under AGV startup conditions. IEC Design N and Design H classifications are roughly equivalent to NEMA Design B and C, respectivement [8].

Pour les moteurs BLDC utilisés dans les AGV, the NEMA design classification is less directly applicable because BLDC motors are electronically commutated and their torque-speed curve is determined by the controller, not the rotor design. Cependant, the concept of locked-rotor (départ) torque maps to the BLDC motor’s peak torque rating, which is typically 2–3× the continuous torque rating for 30–60 seconds before thermal protection activates.

The relationship between NEMA torque classifications and BLDC motor selection is discussed in our BLDC motor vs. servomoteur comparaison, which covers how electronic commutation changes the torque-speed envelope.

Pic vs. Couple continu: Why Most Sizing Errors Happen Here

The single most common mistake in AGV motor selection is sizing for peak torque without validating continuous thermal performance. Motor datasheets advertise peak torque prominently because it is the higher number, but peak torque is only available for a limited duration (typically 30–60 seconds) before the winding reaches its thermal limit.

Maxon specifies that motor constants have tolerances of up to ±10% and change with motor temperature — catalog values apply at 25°C, and a warm motor produces less torque [11]. Faulhaber’s DC Motors Technical Information notes that for optimal motor operation, the required speed should be higher than half the no-load speed, and the load torque should be less than the maximum continuous torque [12]. Yaskawa’s SigmaSelect sizing software generates a comparison report between servo system capability and application requirements, explicitly separating peak and continuous operating points [13].

Practical rule: Size the continuous torque to cover the RMS torque of the duty cycle (including derating for ambient temperature), then verify that the peak torque covers the worst-case transient. If the peak/continuous ratio of your selected motor is less than 2.0×, you may need a larger motor even if the continuous torque appears adequate.

Gearbox Matching: Reflected Torque and Inertia

The gearbox does more than reduce speed and multiply torque. It also transforms the load inertia as seen by the motor, which affects control stability and acceleration response.

Inertie réfléchie

The load inertia reflected to the motor shaft is divided by the square of the gearbox ratio:

J_reflected = J_charger / je²

Où J_charger is the vehicle inertia at the wheel and je is the gearbox ratio. UN 20:1 gearbox reduces the reflected inertia by a factor of 400, making the motor see a much smaller inertia. This is critical for servo-controlled AGVs where the inertia ratio affects tuning stability.

For AGV applications using BLDC motors with Hall-sensor feedback, an inertia ratio below 10:1 is generally acceptable because the velocity control loop does not require the precision of a position loop. For applications requiring precise positioning (par ex., AGV docking), consider upgrading to a motor controller with encoder feedback and targeting an inertia ratio below 5:1.

Gearbox Selection for AGV Drives

For AGV drive systems, the gearbox ratio is selected to place the motor’s operating speed in its efficiency sweet spot (typically 1,500–3,000 RPM for BLDC motors). Below 1,000 RPM, torque ripple and cogging become noticeable; above 3,500 RPM, bearing life degrades and noise increases [12].

Notre spur vs. planetary gearbox comparison provides a detailed analysis of why planetary gearboxes are preferred for AGV applications — higher torque density, lower backlash, and coaxial output that simplifies wheel-hub integration. For applications requiring right-angle output (par ex., steering drives), notre gearbox product page lists NMRV worm gearboxes and bevel-helical options.

Traction Verification: Preventing Wheel Slip

Calculating the required torque is necessary but not sufficient. The torque must be transmissible through the wheel-floor contact. If the applied torque exceeds the friction limit, the wheel slips — and encoder feedback becomes unreliable, directly affecting navigation accuracy [1].

Traction Limit Formula

F_{traction_max} = μ × N_conduire

Where μ is the static friction coefficient between wheel and floor, and N_conduire is the normal force on the driven wheel (not the total vehicle weight — only the weight borne by the driven wheels).

Verification rule: The per-wheel tractive force (F_total / n_conduire) must not exceed F_{traction_max}. If it does, either increase the number of driven wheels, add ballast to increase normal force on driven wheels, or select a higher-friction tire compound.

Pour AGV à entraînement différentiel, the in-place rotation condition is the most likely to cause slip because all the tractive force is concentrated on two wheels pivoting in place. If the calculated F_rotation exceeds the traction limit, the AGV will scrub instead of rotating cleanly, causing tire wear and position error.

Reading Motor Datasheets: Maxon, Faulhaber, et Yaskawa

Motor manufacturers present torque data in different formats. Understanding how to read these datasheets is essential for accurate AGV motor selection.

Maxon: Speed-Torque Line and Motor Constants

Maxon’s catalog specifies the speed-torque line, which is linear for coreless DC motors and BLDC motors with slotless windings. The key parameters are [11]:

• Constante de couple (k_M) in mNm/A — the proportional relationship between current and torque. For coreless Maxon motors, torque and current are strictly proportional, allowing the motor to function as a torque sensor by measuring current.
• Speed-torque gradient (Δn/ΔM) in rpm/mNm — how much speed drops per unit of torque increase. A smaller value means a stiffer motor. The gradient is constant for most motors and equals the ratio of no-load speed to stall torque.
• Nominal torque (maximum continuous torque) — the torque the motor can deliver indefinitely at 25°C ambient without exceeding its thermal class.
• Couple de stand — the torque at zero speed. Never an operating point; causes rapid overheating.

Maxon notes that motor constants have tolerances of up to ±10% and change with temperature. A warm motor is weaker — the speed-torque gradient increases as the motor heats up. This means a motor sized at the edge of its continuous torque rating at 25°C may be underpowered at 40°C ambient.

Faulhaber: Operating Range and Thermal Limits

Faulhaber’s DC Motors Technical Information [12] defines the motor’s operating range on the speed-torque diagram, bounded by:

• Maximum continuous torque (thermal limit line) — the torque sustainable indefinitely.
• Vitesse maximum (mechanical limit) — determined by bearing and commutation capabilities.
• Maximum output power line — typically at 50% of stall torque and 50% of no-load speed.

Faulhaber recommends operating the motor such that the required speed is higher than half the no-load speed at nominal voltage, and the load torque is less than the maximum continuous torque. This ensures the motor operates in its efficient range and avoids excessive copper losses.

Yaskawa: SigmaSelect Sizing Methodology

Yaskawa’s SigmaSelect software [13] takes a system-level approach to servo motor selection. The user inputs:

• Application load data (masse, friction, external forces)
• Mechanical transmission parameters (gearbox ratio, efficacité, inertie)
• Motion profile (vitesse, accélération, dwell time)

The software then generates a report comparing the servo system’s capability (couple maximal, couple continu, vitesse, thermal capacity) against the application’s requirements (RMS torque, couple maximal, vitesse maximum). This report format is valuable because it explicitly separates peak and continuous operating points and includes a thermal margin calculation. While Yaskawa’s SigmaSelect is designed for AC servo motors, the methodology applies directly to BLDC servo systems used in AGVs.

Seven Common Torque Sizing Mistakes

6-Step Torque Selection Workflow

The following workflow consolidates the methodology from this guide into a practical sequence for AGV motor selection. It is compatible with the approaches used by Oriental Motor’s AGV sizing tool [3], iNetic Motion’s calculator [4], and Yaskawa’s SigmaSelect [13].

For AGV applications requiring precise positioning (docking, pallet handling), also evaluate the servo motor vs. moteur pas à pas tradeoff, and consider the direct drive vs. motoréducteur comparison for hub-drive configurations. Notre BLDC contre. servo motors for AGVs analysis provides a three-layer comparison (standard BLDC, Servomoteur BLDC, Servomoteur CA) specific to AGV drive systems.

FAQ

How much torque does a typical AGV need?

It depends on payload. UN 150 kg AMR needs approximately 0.4–0.7 N·m continuous motor torque per wheel (with a 20:1 boîte de vitesses). UN 1,200 kg AGV needs 3–5 N·m. A 5-ton transfer cart needs 10–15 N·m. The quick-reference table in Section 2 provides values for five payload classes.

Quelle est la formule du couple du moteur AGV?

T_moteur = (F_total ×r_roue × SF) / (n_conduire × je × n_g), où F_total =F_rouler + F_acc + F_grade, SF is the safety factor, i is the gearbox ratio, and η_g est l'efficacité de la boîte de vitesses.

What safety factor should I use for AGV torque?

1.25–1.5 for indoor AMRs on smooth floors. 1.5–2.0 for industrial AGVs with ramps or frequent starts. 2.5 for safety-critical applications (médical, préparation des aliments, molten metal transport). The safety factor covers measurement uncertainty, friction variation, and degradation over the motor’s service life.

How does slope angle affect AGV torque?

Grade resistance is F_grade = m × g × péché(je). UN 3% ramp (1.72°) adds 294 N per 1,000 kg of mass. Pour un 1,200 kg AGV, the grade force on a 3% ramp equals 353 N — nearly matching the rolling resistance. Always size for the worst-case slope in the operating environment.

Should I size for peak or continuous torque?

Both. The continuous torque must exceed the RMS torque of the duty cycle (validation thermique). The peak torque must exceed the worst-case transient (acceleration from standstill, ramp start, in-place rotation). Peak torque is typically 2–3× continuous for BLDC motors, available for 30–60 seconds.

What IEC standard applies to AGV motor torque?

CEI 60034-1:2022 defines duty cycle classifications (S1 à S10). Most AGVs operate under S3 (intermittent periodic) ou S4 (intermittent with starting influence). The motor’s rated torque must exceed the RMS torque at the operating ambient temperature, accounting for thermal derating per IEC 60034-1 thermal class limits.

Prochaines étapes

If you have your AGV parameters ready — gross mass, vitesse cible, accélération, pente, wheel diameter — our engineering team can run the torque calculation and recommend a motor, boîte de vitesses, and controller combination. Contacter GreenSky Power with your specifications, or browse our complete motor product catalog for BLDC motors, réducteurs planétaires, et contrôleurs de moteur suitable for AGV drive systems.

All GreenSky Power motors are tested per CEI 60034 et GB/T 1032 testing standards, with dynamometer test reports included with every shipment. For AGV-specific applications, nous offrons conception de moteur sur mesure with integrated encoder, frein, and gearbox options.

Références

1. Honest Edrive Equipment Co., Ltd. (2026). Couple, Traction, and Tread: Engineering Factors in AGV Drive Wheels. Récupéré de HTTPS://www.hagvwheel.com/engineering-factors-in-agv-drive-wheels.html
2. Roue motrice AGV. (2026). Comment calculer le couple de la roue motrice de l'AGV et le dimensionnement du moteur. Récupéré de HTTPS://agvdrivewheel.com/blog/how-to-calculate-agv-drive-wheel-torque-and-motor-sizing
3. Moteur oriental. (2026). AGV — Outil de dimensionnement automatique des véhicules. Récupéré de HTTPS://www.orientalmotor.com/motor-sizing/agv-sizing.html
4. iNetic Motion. (2026). VAG & AMR Motor Calculator for Robotics and Mobility. Récupéré de HTTPS://ineticmotion.com/agv-motor-calculator/
5. Équipement intelligent de Yikong (Bicontrôles). (2026). Guide de dimensionnement du moteur AGV à roue motrice différentielle: Calcul du couple et adaptation de l'inertie. Récupéré de HTTPS://en.bicontrols.com/news_detail/104.html
6. Équipement intelligent de Yikong (Bicontrôles). (2025). Torque Calculation and Optimization for AGV Drive Motors: Enabling Flexible Logistics in Automotive Manufacturing. Récupéré de HTTPS://en.bicontrols.com/news_detail/50.html
7. Commission électrotechnique internationale. (2026). CEI 60034-1:2022 — Rotating Electrical Machines — Part 1: Notation et performances. Edition 15. Genève: CEI. Récupéré de https://www.iec.ch/government-regulators/electric-motors
8. Engineering ToolBox. (2026). Electric Motors — IEC and NEMA Standard Torques. Récupéré de HTTPS://www.engineeringtoolbox.com/iec-nema-standards-torques-d_741.html
9. Siddiqui, F. UN., et autres. (2022). “Motor Parametric Calculations for Robot Locomotion.” Actes d'ingénierie, 20(1), 8. Mdpi. Récupéré de HTTPS://www.mdpi.com/2673-4591/20/1/8
10. DFRobot. (2025). How to Calculate the Motor Torque for a Mobile Robot. Récupéré de HTTPS://wiki.dfrobot.com/tutorial/20135
11. Groupe Maxon. (2025). Motor Data and Simulation — maxon Support. Standard Specification 100 (Docteur moteur) / 101 (EC Motor). Récupéré de HTTPS://support.maxongroup.com/hc/en-us/articles/360013761160-Motor-data-and-simulation
12. Docteur. Fritz Faulhaber GmbH & Co. KG. (2022). Technical Information: Moteurs à courant continu. Récupéré de HTTPS://www.faulhaber.com/fileadmin/Import/Media/EN_TECHNICAL_INFORMATION.pdf
13. Yaskawa Amérique, Inc.. (2025). SigmaSelect Servo Sizing Software — Product Overview. Récupéré de HTTPS://www.yaskawa.com/products/motion/sigma-7-servo-products/software-tools/sigmaselect
14. University of Florida, Machine Design Lab. (2015). Useful Motor/Torque Equations for EML2322L. Récupéré de HTTPS://web.mae.ufl.edu/designlab/motors/Useful%20Equations.pdf
15. AGV Motor. (2025). AGV/AMR Design Calculator: Key Points from Parameter Calculations to Selection Guidelines. Récupéré de HTTPS://agvmotor.com/blogs/knowledge/agv-amr-design-calculator-key-points-from-parameter-calculations-to-selection-guidelines