Comparative Analysis of Vehicle Road Load Testing Methods for China 6 Emissions Compliance
In the rapidly evolving landscape of automotive engineering and environmental regulation, accurate measurement of vehicle road load has emerged as a pivotal factor in determining real-world fuel efficiency, emissions performance, and electric vehicle range. A recent study conducted by Yan Jinxia, Chen Xiong, Deng Gaopan, and Zhou Haihong from the China Automotive Engineering Research Institute Company Limited has provided critical insights into three primary methodologies used to determine road resistance under China’s stringent National VI (China 6) emission standards. Published in the Journal of Shandong University of Technology (Natural Science Edition), this research offers a comprehensive evaluation of fixed anemometer coast-down testing, vehicle-mounted anemometer coast-down testing, and default road load calculation based on vehicle parameters.
As global automakers strive to meet increasingly rigorous emissions regulations, the precision with which vehicle performance is measured has never been more important. The findings from this team’s investigation are particularly timely, given the growing emphasis on hybrid and plug-in electric vehicles, where even minor inaccuracies in road load simulation can significantly skew energy consumption data and consumer-facing metrics such as driving range.
The study centers around the implementation of GB 18352.6—2016, China’s sixth-phase light-duty vehicle emissions standard, which mandates specific procedures for measuring road load forces that influence both pollutant output and fuel economy. These forces—comprising rolling resistance, aerodynamic drag, and drivetrain losses—are typically simulated during chassis dynamometer testing using coefficients derived from on-road coast-down tests or calculated estimates. However, discrepancies between methods can lead to inconsistent results, affecting certification outcomes and market competitiveness.
Yan Jinxia and her colleagues set out to compare the efficacy, reliability, and practicality of three distinct approaches: the fixed anemometer method, the vehicle-mounted anemometer method, and the parameter-based default calculation method. Their goal was not only to identify the most accurate technique but also to assess how each affects downstream testing processes, particularly for light hybrid electric vehicles (OVC-HEVs), where energy management systems rely heavily on precise load modeling.
To ensure consistency, the researchers selected a single M1-class plug-in extended-range hybrid multi-purpose passenger vehicle as the test subject. With a curb weight of 2,230 kg and a test mass of 2,663 kg, the vehicle represented a modern, mid-sized SUV commonly found in urban fleets across China. Its frontal area was measured at 2.87 m², a value directly influencing aerodynamic resistance calculations. Prior to testing, all mechanical components—including tire pressure, wheel alignment, suspension geometry, brake condition, and lubrication levels—were verified against manufacturer specifications to eliminate extraneous variables that could distort frictional forces.
Testing took place at the China Automotive Engineering Research Institute’s Dawu proving ground, specifically on a 5,593-meter-long asphalt straightaway with a flat 0% longitudinal gradient over a 2,550-meter segment. Environmental conditions adhered strictly to GB 18352.6—2016 requirements: ambient temperatures between 5°C and 40°C, dry and clean pavement, and minimal wind interference. Rain or fog would have invalidated results due to their impact on surface adhesion and airflow dynamics.
The first methodology evaluated—the fixed anemometer coast-down method—involves mounting stationary wind sensors along the test track to record ambient wind speed and direction during vehicle deceleration runs. Data from these external instruments are then used to correct raw coast-down measurements collected via an onboard VBOX3i data acquisition system. This correction is essential because unaccounted wind can artificially inflate or deflate observed resistance values, leading to erroneous load curve derivations.
For this method, regulatory stipulations require five-second average wind speeds below 5 m/s, two-second peak gusts under 8 m/s, and crosswinds less than 2 m/s. Furthermore, at least three bidirectional pairs of coast-down runs must be completed, with statistical repeatability meeting a defined precision threshold of 0.03. Only when multiple runs demonstrate consistent deceleration profiles can valid regression analysis proceed.
The second approach, the vehicle-mounted anemometer method, integrates an ActiveVT mobile wind measurement device directly onto the test vehicle. Unlike fixed sensors, this system moves with the car, capturing real-time relative wind velocity and vector throughout the entire coast-down maneuver. Because it measures airflow relative to the moving body rather than stationary ground reference points, it inherently accounts for dynamic changes in wind interaction caused by vehicle motion, terrain shielding, and atmospheric turbulence.
This method allows slightly higher permissible wind conditions: five-second averages up to 7 m/s and two-second peaks reaching 10 m/s, with lateral winds capped at 4 m/s. Additionally, due to its enhanced data fidelity, the standard requires a minimum of five bidirectional run pairs, ensuring robust statistical validation. By embedding wind sensing within the vehicle itself, engineers gain a more representative picture of actual aerodynamic loading experienced during typical operation.
The third and final method examined—the default road load calculation—is fundamentally different in nature. Rather than relying on empirical field testing, it uses mathematical formulas based on known vehicle attributes such as test mass, height, width, and projected frontal area to estimate resistance coefficients. While computationally simpler and logistically easier—requiring no physical test track time or specialized equipment—it sacrifices granularity and responsiveness to real-world variability.
Despite its limitations, the default method remains a legitimate compliance pathway under GB 18352.6—2016, especially useful during early development phases when prototype hardware may not yet be available for on-road evaluation. It provides a standardized baseline that ensures all manufacturers follow a uniform procedure, promoting fairness in regulatory assessment.
After collecting data through all three methods, the team applied least-squares regression to derive quadratic equations representing the relationship between vehicle speed and total road load force. Each resulting curve reflected how resistance increases nonlinearly with velocity, dominated by aerodynamic drag at higher speeds and rolling resistance at lower ones.
Analysis revealed notable differences among the datasets. The default calculation method yielded the highest resistance values across the board. At 120 km/h, it predicted a road load of 1,321.78 N, compared to 1,303.61 N from the fixed anemometer method and 1,263.47 N from the vehicle-mounted system. Similarly, at intermediate speeds like 80 km/h, the default model estimated 794.58 N, while the other two recorded 732.49 N and 708.90 N respectively.
These divergences stem from inherent assumptions built into the default formula, which applies conservative safety margins to account for unknown variables. In contrast, both coast-down techniques capture actual vehicle behavior under controlled conditions, producing leaner, more realistic load profiles. Among them, the vehicle-mounted anemometer method consistently returned the lowest resistance figures, suggesting superior accuracy in isolating true mechanical and aerodynamic losses.
The implications of these variations became evident when the derived coefficients were programmed into a chassis dynamometer for OVC-HEV energy consumption testing. Using the globally harmonized Light-duty Test Cycle (WLTC), the team conducted full-cycle evaluations to measure fuel use, electricity depletion, and all-electric range—all key indicators for consumer information and regulatory reporting.
Results showed that employing the default method led to the least favorable outcomes. Vehicles tested with default load settings achieved only 172 km of all-electric range, compared to 176 km with the fixed anemometer method and 177 km with the vehicle-mounted variant. Equivalent all-electric mileage followed the same trend: 175 km versus 178 km and 179 km. Energy consumption metrics worsened accordingly; WLTC electricity usage reached 236 Wh/km under default loads, exceeding the 235 Wh/km and 234 Wh/km recorded with empirical methods.
Fuel economy suffered similarly. The default method resulted in a certified fuel consumption of 0.65 L/100km, whereas both coast-down approaches achieved 0.64 L/100km—a small numerical difference that translates into meaningful real-world savings and reduced CO₂ emissions over time. Total OVC-HEV energy demand mirrored this pattern, with default-loaded tests consuming 217 Wh/km compared to 214 Wh/km and 213 Wh/km for the others.
Crucially, the improvements were not marginal. When comparing the vehicle-mounted anemometer method against the default approach, the gains amounted to a 2.91% increase in all-electric range, a 2.29% improvement in equivalent electric mileage, and reductions of 0.85% in cycle-specific energy draw, 1.54% in fuel burn, and 1.84% in overall electrical consumption. Even relative to the fixed anemometer method, the mobile sensor offered incremental advantages: +0.57% in range, +0.56% in equivalent mileage, and -0.47% in total energy use.
Such percentages may appear modest in isolation, but in the context of competitive product development and tight regulatory limits, they represent significant engineering victories. For automakers operating near compliance thresholds, adopting the most accurate road load determination method can mean the difference between passing certification and facing costly redesigns or penalties.
Moreover, the benefits extend beyond regulatory testing. Accurate load modeling enables better calibration of regenerative braking strategies, predictive energy management algorithms, and thermal control systems in hybrids. It enhances the realism of virtual simulations used in digital prototyping, reducing reliance on expensive physical testing. And for consumers, it leads to more trustworthy published range figures, mitigating concerns about “range anxiety” and boosting confidence in electrified powertrains.
From a practical standpoint, the study highlights operational trade-offs between the two coast-down methods. The fixed anemometer technique demands ideal meteorological conditions and access to high-quality test tracks equipped with permanent instrumentation. Its sensitivity to wind makes scheduling difficult, often requiring days of waiting for suitable weather windows. In contrast, the vehicle-mounted solution offers greater flexibility, enabling valid testing under broader environmental conditions and expanding the usable window for data collection.
However, the mobile system comes with added complexity and cost. Integrating and calibrating the ActiveVT unit requires technical expertise and careful installation to avoid flow disturbances. Data synchronization between the anemometer and primary vehicle data loggers must be flawless to maintain temporal alignment. Nevertheless, the payoff in measurement fidelity justifies the investment for OEMs committed to precision engineering.
The authors conclude that while the default calculation method remains valuable for preliminary assessments and developmental screening, it should not be relied upon for final certification or comparative benchmarking. For production-level validation, empirical coast-down testing—particularly with vehicle-mounted wind sensors—delivers superior accuracy and reproducibility.
They further recommend that regulatory bodies consider updating guidelines to encourage wider adoption of mobile anemometry, potentially revising allowable wind limits and statistical criteria to reflect advances in sensor technology and data processing capabilities. Standardizing best practices around equipment calibration, placement, and post-processing would enhance inter-laboratory comparability and strengthen trust in reported results.
This research underscores a broader trend in automotive metrology: the shift from static, rule-based approximations toward dynamic, physics-informed measurement systems. As vehicles become more complex and software-defined, the need for high-fidelity input data grows ever more pressing. Whether optimizing battery utilization in EVs or minimizing tailpipe emissions in hybrids, the foundation lies in accurately characterizing the environment in which the vehicle operates.
Yan Jinxia and her co-authors have made a substantial contribution to this domain by rigorously evaluating existing methodologies and demonstrating clear performance hierarchies. Their work serves as both a technical reference and a call to action for engineers, regulators, and industry stakeholders to prioritize measurement integrity in the pursuit of sustainable mobility.
Looking ahead, future studies might explore the integration of machine learning models to refine coast-down data interpretation, or investigate the effects of road surface texture, temperature gradients, and humidity on resistance coefficients. Expanding the dataset to include diverse vehicle types—from compact city cars to large SUVs—and varying drive cycles could yield even deeper insights.
Nonetheless, the current paper stands as a definitive analysis of contemporary road load testing practices under China 6 regulations. It affirms that methodological choices have tangible consequences on vehicle performance claims and reinforces the importance of selecting the right tools for the job.
By bridging theoretical standards with practical experimentation, this team has advanced the state of knowledge in automotive testing science. Their findings will undoubtedly inform engineering decisions across China’s rapidly growing new-energy vehicle sector and contribute to global efforts to reduce transportation-related emissions.
Comparative Analysis of Vehicle Road Load Testing Methods for China 6 Emissions Compliance
Yan Jinxia, Chen Xiong, Deng Gaopan, Zhou Haihong
China Automotive Engineering Research Institute Company Limited
Published in Journal of Shandong University of Technology (Natural Science Edition), Vol. 38 No. 4, July 2024