An optimal gear-shifting strategy for heavy trucks with trade-off study between trip time and fuel economy

Xinxin Zhao, Adam H. Ing, Nasser L. Azad, John McPheeInternational Journal of Heavy Vehicle Systems 2015 · DOI

TL;DR

This paper presents an optimal gear-shifting strategy for heavy mining trucks to simultaneously minimize trip time and fuel consumption. By developing both high-fidelity and low-order models of the truck and employing dynamic programming, they demonstrate a potential annual fuel saving of 40,000 liters for a typical mine.

What Problem Does This Paper Address?

Heavy mining trucks are crucial for moving large payloads but are significant fuel consumers. While hybrid powertrains offer fuel efficiency gains, their cost is prohibitive for fleet replacement. This paper tackles the problem of improving fuel efficiency and reducing operational costs by optimizing the existing gear-shifting strategy, a more cost-effective approach. The challenge lies in balancing two competing objectives: minimizing trip time (for operational efficiency) and minimizing fuel consumption (for cost savings and environmental impact).

Methods

The authors developed two models of the mining truck:

  • High-fidelity model: A detailed, quasi-static model built in MapleSim, capturing complex dynamics like engine combustion, turbocharging, and a multi-planetary gear transmission. This model is computationally intensive but provides a realistic representation.
  • Low-order model: A steady-state, backwards model developed in Matlab. This model is computationally efficient and suitable for control design, though it simplifies some dynamics like inertial effects.

The core of the optimization was performed using dynamic programming. This technique was applied to the low-order model to find an optimal control law for gear shifting over a defined route. An objective function was formulated to combine fuel consumption and trip time using weighting coefficients ($\lambda_1$ for fuel, $\lambda_2$ for time). A Pareto curve was generated to visualize the trade-off between these two objectives for different weighting combinations. Finally, the optimal strategy derived from the low-order model was tested on the high-fidelity model to validate its performance.

Datasets

The paper utilizes:

  • Experimental engine data: Used to build the fuel consumption map for the diesel engine in the high-fidelity model.
  • Road topography data: Gathered using Google Earth for a specific 3.4 km route in a North-Western China colliery. This data includes altitude and slope information, which is crucial for the optimization.

No new benchmarks were introduced; the focus was on applying existing modeling and optimization techniques to a specific heavy-duty vehicle application.

Key Insights and Ideas

  • Trade-off is inherent: There's no single "best" gear-shifting strategy; it's always a compromise between getting there fast and using less fuel. The Pareto curve clearly illustrates this.
  • Mixed weighting is key: While weighting solely for fuel consumption (λ₁=1, λ₂=0) might seem intuitive for maximum savings, the study found that a mixed weighting (e.g., λ₁=0.4, λ₂=0.6) can achieve lower fuel consumption and better overall efficiency by keeping the engine in its most efficient operating region.
  • Low-order models are powerful for control design: The low-order model, despite its simplifications, was sufficient to derive an optimal control strategy that performed well when applied to the complex high-fidelity model. This highlights the utility of control-oriented modeling.
  • Real-world route matters: Optimizing for a specific, known route with its unique topography is more effective than relying on generic driving cycles for heavy-duty trucks.

Key Results

  • The developed optimal gear-shifting strategy, when applied to the high-fidelity model, showed a reduction in trip time by over 1.7% and fuel consumption by approximately 1% compared to a rule-based current strategy.
  • For a mine producing 110 kilotonnes of coal per day, using a 30-tonne truck, this strategy could lead to an estimated annual fuel saving of 40,000 liters of diesel.
  • The low-order and high-fidelity models showed consistent results in terms of trip time and fuel consumption, validating the approach. The low-order model predicted a slightly longer trip time (294s vs. 291.3s) and slightly higher fuel consumption (3.27L vs. 3.21L) compared to the high-fidelity model, with an error of about 1.8%.

Limitations and Future Work

  • Offline optimization: The current strategy is determined offline based on a known route. The authors suggest that future work could focus on developing an online optimal controller that can adapt to changing road conditions in real-time.
  • Experimental validation: The study relies on simulations. The authors explicitly state that future work should validate these results against experimental data from a real vehicle.
  • Model simplifications: While the low-order model is effective for control design, it ignores inertial effects. The high-fidelity model, though detailed, is quasi-static. More dynamic models could potentially yield further improvements.
  • Generalizability: The optimization was performed for a specific route. Testing the strategy on a wider variety of routes and conditions would be beneficial.

Summary updated

Abstract

We show how the fuel efficiency of heavy mining trucks can be improved by optimising the gear-shifting strategy. Using characteristic tests of the diesel engine, a high-fidelity model of a mining truck was built in MapleSim and a consistent low-order model was developed in Matlab. Dynamic programming was used to optimise the low-order model of the specialised off-road 30-tonne truck over a fixed route in a mining area. There were two competing objectives: fuel use and trip time, which were combined in a single objective function using weighting coefficients. A Pareto curve was created to analyse the effect of the weights on the fuel use and trip time. Applying the control strategy obtained from dynamic programming to the high-fidelity model, it is estimated that 40,000 L of fuel can be saved annually for a mine that produces 110 kilotonnes of coal per day.
BibTeX
@article{zhao2015optimal,
  author = {Xinxin Zhao and Adam H. Ing and Nasser L. Azad and John McPhee},
  title = {{An optimal gear-shifting strategy for heavy trucks with trade-off study between trip time and fuel economy}},
  journal = {International Journal of Heavy Vehicle Systems},
  year = {2015},
  doi = {10.1504/ijhvs.2015.073205},
}
Figure