Advanced

On Optimal Control for Concept Evaluation and System Development in Construction Machines

Frank, Bobbie LU (2018)
Abstract
The main goal with this thesis is to develop a method to maximize the fuel efficiency [ton/l] in construction machines, while fulfilling the desired productivity [ton/h]. This is achieved by focusing on two of the main influencers; the machine concept evaluation and the development of operator assist systems. To be able to perform a concept evaluation that is unbiased from control engineering experience and test repetitiveness, an optimal control algorithm based on dynamic programming, which ensures global optimum, is developed. This algorithm is also able to handle the system optimization that is a necessity when performing a machine concept evaluation. The optimal control results from the concept evaluation are able to provide input when... (More)
The main goal with this thesis is to develop a method to maximize the fuel efficiency [ton/l] in construction machines, while fulfilling the desired productivity [ton/h]. This is achieved by focusing on two of the main influencers; the machine concept evaluation and the development of operator assist systems. To be able to perform a concept evaluation that is unbiased from control engineering experience and test repetitiveness, an optimal control algorithm based on dynamic programming, which ensures global optimum, is developed. This algorithm is also able to handle the system optimization that is a necessity when performing a machine concept evaluation. The optimal control results from the concept evaluation are able to provide input when developing control strategies for operator assist systems, automatic functions and autonomous machine control.
The method is demonstrated on a wheel loader, working in a production chain, but can be applied to other construction machines, for example articulated haulers and excavators. Common denominators can be found with forestry equipment, agriculture machines and on road vehicles. The optimal control algorithm is put to test by challenging the calculated theoretical optimum with measured data from an extensive empirical study to test the validity of the global optimum found.
The optimal control algorithm, based on dynamic programming, successfully works. The result demonstrates approximately 14% better fuel efficiency in a gravel application, compared to the most fuel efficient operator’s best work cycle in the empirical study. The best measured work cycle is approximately 30% better than the average in the study. A similar result is shown in a timber grapple application.
The proposed method, and the algorithm developed, works for all three investigated machine concepts, enabling an unbiased concept evaluation and system optimization. The method is demonstrated on a concept comparison between a conventional wheel loader, a parallel hybrid wheel loader and a series hybrid wheel loader. The results from the concept comparison example indicate that the parallel hybrid has about 5% higher fuel efficiency and the series hybrid wheel loader has around 23% higher fuel efficiency compared to the conventional machine, which is used as baseline, at the same productivity. In an example of a system optimization for the primary energy converter, the genset, internal combustion engine and electrical machine, in the series hybrid wheel loader indicate that the optimal power rating of the genset in the investigated application is 0.6 times the internal combustion engine power rating of the conventional wheel loader. Even if the factor 0.5 showed even higher fuel efficiency, that power rating is not allowed due to complete machine performance requirements. The difference between the largest genset power rating, which is equivalent to the power rating of the conventional wheel loader, and the optimal, is approximately 6% higher fuel efficiency at the same productivity.
With the proposed method it is shown how to extract input to the development of operator assist systems, automatic functions, and autonomous construction machine control development from the optimal control results. The results are attained from the optimal control calculations performed in early development in the concept evaluation and system optimization. This also implies that the optimal control results, if used in the development of these advanced functions can increase the average fuel efficiency by up to 35 45%. The percentage is dependent on the operator´s proficiency and application according to the conclusions from the empirical study. A suggestion of how to use the optimal control results as input, when developing operator assist systems, automatic functions and autonomous machine control, is also presented. (Less)
Please use this url to cite or link to this publication:
author
supervisor
opponent
  • Professor Rizzoni, Giorgio, Ohio State University, USA
organization
publishing date
type
Thesis
publication status
published
subject
keywords
Construction Machines, Wheel Loader, Fuel Efficiency, Productivity, Operator Behavior, Optimal Control, Dynamic Programming, Complete Work Cycle, Actuator Movement Optimization, Bucket Fill Trajectory, Gravel Simulation, Concept Evaluation, System Optimization, Operator Assist Systems, Automatic Functions, Autonomous Machines.
pages
134 pages
publisher
Department of Biomedical Engineering, Lund university
defense location
lecture hall M:B, building M, Ole Römers väg 1, Lund University, Faculty of Engineering LTH, Lund
defense date
2018-09-07 11:15
ISBN
978-91-88934-90-1
978-91-88934-91-8
language
English
LU publication?
yes
id
a1536319-4cf9-4ec7-aef7-09112536b0c8
date added to LUP
2018-06-13 14:50:16
date last changed
2019-02-01 15:34:18
@phdthesis{a1536319-4cf9-4ec7-aef7-09112536b0c8,
  abstract     = {The main goal with this thesis is to develop a method to maximize the fuel efficiency [ton/l] in construction machines, while fulfilling the desired productivity [ton/h]. This is achieved by focusing on two of the main influencers; the machine concept evaluation and the development of operator assist systems. To be able to perform a concept evaluation that is unbiased from control engineering experience and test repetitiveness, an optimal control algorithm based on dynamic programming, which ensures global optimum, is developed. This algorithm is also able to handle the system optimization that is a necessity when performing a machine concept evaluation. The optimal control results from the concept evaluation are able to provide input when developing control strategies for operator assist systems, automatic functions and autonomous machine control.<br/>The method is demonstrated on a wheel loader, working in a production chain, but can be applied to other construction machines, for example articulated haulers and excavators. Common denominators can be found with forestry equipment, agriculture machines and on road vehicles. The optimal control algorithm is put to test by challenging the calculated theoretical optimum with measured data from an extensive empirical study to test the validity of the global optimum found.<br/>The optimal control algorithm, based on dynamic programming, successfully works. The result demonstrates approximately 14% better fuel efficiency in a gravel application, compared to the most fuel efficient operator’s best work cycle in the empirical study. The best measured work cycle is approximately 30% better than the average in the study. A similar result is shown in a timber grapple application.<br/>The proposed method, and the algorithm developed, works for all three investigated machine concepts, enabling an unbiased concept evaluation and system optimization. The method is demonstrated on a concept comparison between a conventional wheel loader, a parallel hybrid wheel loader and a series hybrid wheel loader. The results from the concept comparison example indicate that the parallel hybrid has about 5% higher fuel efficiency and the series hybrid wheel loader has around 23% higher fuel efficiency compared to the conventional machine, which is used as baseline, at the same productivity. In an example of a system optimization for the primary energy converter, the genset, internal combustion engine and electrical machine, in the series hybrid wheel loader indicate that the optimal power rating of the genset in the investigated application is 0.6 times the internal combustion engine power rating of the conventional wheel loader. Even if the factor 0.5 showed even higher fuel efficiency, that power rating is not allowed due to complete machine performance requirements. The difference between the largest genset power rating, which is equivalent to the power rating of the conventional wheel loader, and the optimal, is approximately 6% higher fuel efficiency at the same productivity.<br/>With the proposed method it is shown how to extract input to the development of operator assist systems, automatic functions, and autonomous construction machine control development from the optimal control results. The results are attained from the optimal control calculations performed in early development in the concept evaluation and system optimization. This also implies that the optimal control results, if used in the development of these advanced functions can increase the average fuel efficiency by up to 35 45%. The percentage is dependent on the operator´s proficiency and application according to the conclusions from the empirical study. A suggestion of how to use the optimal control results as input, when developing operator assist systems, automatic functions and autonomous machine control, is also presented.},
  author       = {Frank, Bobbie},
  isbn         = {978-91-88934-90-1},
  keyword      = {Construction Machines, Wheel Loader, Fuel Efficiency, Productivity, Operator Behavior, Optimal Control, Dynamic Programming, Complete Work Cycle, Actuator Movement Optimization, Bucket Fill Trajectory, Gravel Simulation, Concept Evaluation, System Optimization, Operator Assist Systems, Automatic Functions, Autonomous Machines.},
  language     = {eng},
  pages        = {134},
  publisher    = {Department of Biomedical Engineering, Lund university},
  school       = {Lund University},
  title        = {On Optimal Control for Concept Evaluation and System Development in Construction Machines},
  year         = {2018},
}