Purdue University Graduate School
Browse

An Electro-Hydraulic Traction Control System for Heavy Duty Off-Road Vehicles: Formulation and Implementation

Download (35.67 MB)
thesis
posted on 2020-01-16, 19:53 authored by Addison B. AlexanderAddison B. Alexander
Traction control (TC) systems have become quite common in on-road passenger vehicles in recent years. However, for vehicles in other applications, they are not as widely available.
This work presents a methodology for the proper design and implementation of a traction control system for heavy duty off-road machines, using a wheel loader as a reference vehicle.

A simulation model was developed, using standard vehicle dynamics constructs, including equations of motion and a description of the distribution of weight between the axles for different operating conditions. This model contains considerations for resistive forces acting on the machine implement, such as that generated by a work pile. The simulation also incorporates a detailed representation of the slip-friction characteristics between the vehicle tires and the road surface. One objective of this research was to model this interaction accurately, because the system traction behavior is dependent on it. Therefore, a series of tests was run using a state estimator to generate data on the slip-friction relationship at various ground conditions, and the results were incorporated into the simulation model. The dynamics of the machine braking system pressure were also modeled to give a more accurate description of the system response. The result is a mathematical model capable of accurately reproducing the behavior of the real-world system.

One of the primary goals of this work was the description of the traction control strategy itself, which should work as effectively and efficiently as possible. Several different aspects of the system were taken into consideration in generating this control structure. First, a relatively simple controller based on a PID control law was created. This controller was updated to account for peculiarities of the traction control system, as well as aspects like time delay. From there, more advanced controllers were created to address certain aspects of the system in greater detail. First, a self-tuning controller based on real-time optimization strategies was developed, to allow the controller to quickly adapt to changes in ground condition. Then, different nonlinear controllers were synthesized which were designed to address the theoretical behavior of the system. All of these controllers were simulated using the system model and then some were run in experiments to show their potential for improving system performance. To improve system efficiency, the machine drivetrain itself was also examined to develop a more efficient control algorithm. By designing a more efficient methodology, traction control con gurations which had previously seen increases in fuel consumption of 16% were now able to actually reduce fuel usage by 2.6%.

Another main goal of this work was the development of a prototype system capable of implementing the formulated control strategies. The reference machine was modi ed so that the brakes could be controlled electronically and independently for implementation of the TC system. The vehicle was instrumented using a wide array of sensors, and estimation methodologies for accurately determining vehicle speed and implement forces were designed. The velocity estimator designed in this work is more accurate and more reliable than an industry standard sensor, which is important for traction control implementation. The implement force estimate was also quite accurate, achieving payload estimate errors of less than 2.5%, comparable to commercially-available measurement systems. This setup allowed for tests to be accurately compared, to assess the traction control performance.

With the objective of performing experiments on the traction control system, many tests were run to assess its capabilities in various situations. These tests included experiments for characterizing the vehicle behavior so that the simulation model could be updated to accurately reflect the physical machine performance. Another task for the experimental work was the generation of useful metrics for quantifying traction control performance. Laboratory experiments which were very controlled and repeatable were also run for generating data to improve the system model and for comparing traction control performance results side-byside. The test metrics proposed for these experiments provided for accurate, repeatable comparisons of pushing force, tire wear, and brake consumption. For each of these tests, the traction control system saw an increase in pushing force of at least 10% when compared with the stock machine, with certain operating conditions showing increases as high as 60%. Furthermore, every test case showed a decrease in wheel slip of at least 45% (up to 73% for some cases), which translates into increased tire longevity.

Other tests were conducted in the eld, designed to mimic the real-world operating conditions of the wheel loader. Various performance comparisons were made for different con gurations in which traction control could provide potential bene ts. These included parameters for comparing overall vehicle performance in a typical truck loading cycle, such as tire wear, fuel consumption, and material moved per load. Initial results for this testing showed a positive result in terms of wheel slip reduction, but other performance parameters such as fuel consumption were negatively impacted. Therefore, the control structure was reexamined extensively and new methods were added to improve those results. The final control implementation saw a 12% reduction in tire slip, while also reducing fuel consumption by 2.6% compared to the stock system. These results show signi cant potential for traction control as a technology for maximizing the performance output of construction machines.

History

Degree Type

  • Doctor of Philosophy

Department

  • Mechanical Engineering

Campus location

  • West Lafayette

Advisor/Supervisor/Committee Chair

Andrea Vacca

Additional Committee Member 2

Gregory Shaver

Additional Committee Member 3

John Lumkes

Additional Committee Member 4

José Garcia Bravo