Automated Hydraulic System Design and Power Management in Mobile Applications

Since the first oil crisis in the beginning of the 1970'ties there have been an increasing focus on energy and energy consumption, in the latter years also driven by the climate changes that are taking place. Hydraulic systems have, however, traditionally been characterised by low system efficiency and therefore there is today a shift towards using electric drives as replacement for hydraulic drives. There are, however, a number of different areas, where hydraulic systems offers possibilities that cannot be matched by electric drives, as the hydraulic systems are typically characterised by a much higher force, torque and power density. One of these areas is the mobile hydraulic area, which generally comprise all type of off-highway machinery, such as construction equipment, agricultural equipment etc. But where hydraulic systems earlier was designed with primary focus on cost, dynamic performance and accuracy, energy consumption is becoming an ever more important design parameter.

At the same time as the first oil crisis the first hydraulic load sensing (LS) systems also emerged on the market, which, compared to the other systems of the time, offered significant energy saving potentials and which today are found on most medium and high-end mobile hydraulic machinery. Despite the energy saving potentials that these systems posses, compared to the other open-circuit hydraulic system topologies, LS-system may still be subject to very low system efficiencies if not designed correctly. This is typically the case for systems, with highly varying operating conditions and where more work functions (consumers) are operated simultaneously. The low system efficiency is in this regard not necessarily due to low component efficiencies, which often actually have an efficiency comparable to that of electrical machines if operated in the intended and optimal work area, but due to an inappropriate system layout. Most of the power lost in open circuit hydraulic system systems is in this regard in the transmission part, i.e. hoses and fittings, and the valves used to control the system. A large part of the design task is therefore to design the system so these losses may be minimised. The problem with this is that there exist no methods for how to do this, and the design task may generally be a very complex process, which requires substantial prior knowledge and experience. Having designed a systems no methods do either not exist for ensuring that the system designed is actually suited for a given application.

Today a change is furthermore happening, where new and more intelligent components, which are electrically controllable, are emerging and more and more sensors are finding their way into the hydraulic machines. This also means that the door is opened for a new range of possibilities with regard to better system utilisation. The latter is both in regard to new functions and facilities, but also with regard to utilising the system in the most energy optimal way, ensuring that all components are working under the most optimal operating conditions. The above in this way constitute the background for the work that is the basis of this report, which deals with how to design and control open-circuit hydraulic systems with multiple consumers to obtain the largest energy utilization, when also considering other design parameters like installation cost, complexity and system performance.

The report begins with a presentation and definition of the problem considered and a review of the work that has been made within the area of hydraulic load sensing (LS) systems throughout the last three decades. Through this, the different stability problems that are often encountered in LS-systems are explained along with how they may be avoided. In addition hereto an overview of the work that has been made in relation to electronic load sensing (ELS) systems is presented along with an overview of the other energy efficient system topologies that exist. Finally the first part is completed with an overview of the main contributions from the present work, also describing the limitations that are made throughout the project.

As introduction to the actual design approach, the philosophy behind this is then first presented and discussed. The design formulation builds on a multi-level optimisation approach and this is put in perspective to the other approaches that have been attempted, and it is described which informations and what the prerequisites are for the method developed. In this connection the different design criteria are also presented and it is explained how the different work functions may be prioritised relative to each other.

Based on this discussion the developed design framework is presented in terms of the methods that are the foundation of the approach. This first includes a presentation of the used graph theory representation that is developed to represent a hydraulic open-circuit system and which is based on a numerical formulation that uniquely describe the system in terms of five set of design variables that describe respectively the topology, the components and the operating conditions of the system considered. Hereafter it is explained how the data for the different load cases are generated based on a number of motion diagrams for each of the different actuators in the considered system, which is the basis for the design process.

With basis in the established system representation and the load case data, the developed method used for the automated generation and solving of the system equations are then described, where the system equations refers to the set of equations that describe the flow and pressure within the system. The developed method is derived based on ideas from graph theory. Compared to the standard methods the here developed method however utilises another component formulation and it is also expanded to be able to handle special components like differential cylinders and components that have either a zero or an infinite flow conductance. In relation hereto, a number of generic component models are furthermore developed and presented that may be handled by the method, and which follows the standard formulation used.

The method develop is based on a multi-level optimisation approach and from the identified design criteria, corresponding mathematical objective functions are formulated, which are used in the design process. A method for weighting the different criteria relative to each other is furthermore presented. In relation hereto the considerations regarding choice of optimisation method is furthermore discussed. 

Based on the developed design method this used to design a number of test systems as validation of the method. The systems that are designed in this relation is respectively the power supply for a tractor, an example system that is considered a benchmark case and finally the working hydraulics for a forklift truck. For the first and the last application the examples are based on identified duty cycles for the two applications.

Finally a power management algorithm is presented, which is used for the actual control of an open-circuit hydraulic system. This algorithm is based on parts of the aforementioned optimisation procedure, but is extended so the dynamics in the system controlled is also taken into account and so it handles the saturation phenomena that may arise in relation to a hydraulic system, i.e. pressure, flow, power and/or torque saturation. Parts of the pump controller are similar moved to the algorithm, which hereby may be used as both a pump pressure or a pump flow control system. In relation hereto the results of a development work related to designing and controlling an electro-hydraulic actuator for an (electrically controlled) open circuit axial piston pump is presented. The pump in this way constitute part of the prerequisites for the presented power management algorithm. The developed algorithm is afterwards validated through simulation results, where it is shown that it is capable of controlling the system according to the specified controller-criteria and where it at the same time handles the transitions and saturation problems that occur during operation. Practically this is exemplified by controlling the working hydraulics on a backhoe loader, which is also the basis for the description. The developed algorithm is however generally applicable and may be used on all other open-circuit systems that utilises electronically controllable components where there is access to the system operating informations such as system and load pressures.

Finally the main findings of the report is summarised and the areas of further work are described.