energies Article
Reducing Fuel Consumption in Hydraulic Excavators—A Comprehensive Analysis Milos Vukovic, Roland Leifeld * and Hubertus Murrenhoff Institute for Fluid Power Drives and Controls (IFAS), RWTH Aachen University, 52062 Aachen, Germany;
[email protected] (M.V.);
[email protected] (H.M.) * Correspondence:
[email protected]; Tel.: +49-24180-27528 Academic Editor: Andrea Vacca Received: 29 March 2017; Accepted: 9 May 2017; Published: 12 May 2017
Abstract: Mobile machines, especially excavators, still consume considerable amounts of fuel during their operating lifetimes. This is not only undesirable in economic terms but also adversely affects our environment. The following paper discusses methods to lower fuel consumption by conducting a comprehensive analysis of the components comprising a hydraulic excavator and the cycles these machines perform. One of the main aims is to emphasise that a design centred on the standard definitions of efficiency, especially hydraulic efficiency, can be rather misleading. A new approach using a novel fuel consumption model, based on the Willans approximation, coupled with the concepts of fixed and variable fuel consumption is introduced and validated using real test data obtained from an 18 t excavator. The new methodology can be used to help uncover simpler methods to improve today’s machines. Keywords: hydraulic excavators; hybrid systems; fuel consumption; energy recovery; engine downspeeding; independent metering
1. Introduction Hydraulic excavators are responsible for approximately 60% of the CO2 emissions produced by construction machinery [1]. This is in part due to the sheer number of machines in use and also to their extremely low efficiencies of around 10% [2]. Despite their immense impact on our environment many aspects regarding the exact reasons for their high fuel consumption remain misunderstood. The hydraulic systems used to power these machines are often unjustly blamed for the majority of the losses. As a result, much research has gone into the development of more efficient hydraulic architectures capable of lowering so-called throttling losses and enabling energy recovery. Much less attention has been paid to the machine as a whole. The following work aims to clarify many issues and myths surrounding hydraulic excavators by guiding the reader through a comprehensive analysis of the whole machine. The losses occurring in components and subsystems are explained and a detailed discussion of measurement data, obtained from field tests with an 18 t machine, is presented. These thoughts ultimately lead to the introduction of a novel fuel consumption model, capable of describing and predicting the fuel consumption of a machine for all duty cycles. Instead of using the widespread definitions of machine and hydraulic efficiency, the concepts of fixed and variable fuel consumption are introduced. An important aspect of the research is the validation of the methodology using real measurement data. The authors hope to provide a tool that can be used by engineers, during the initial design phase, to easily evaluate the fuel saving potential of different architectures before running complex system simulations. As an example, solutions involving independent metering, displacement control and hybridization are discussed. Along with a brief summary, the paper concludes with an outlook concerning future work. Energies 2017, 10, 687; doi:10.3390/en10050687
www.mdpi.com/journal/energies
Energies 2017, 10, 687
2 of 25
Energies 2017, 10, 687
2 of 25
Energies 2017, 687 2 of 25 2. Review of 10, Components and Subsystems 2. Review of Components and Subsystems Figure 1ofillustrates the typical layout and individual components of a state of the art machine. 2. Review Subsystems Figure 1 Components illustrates theand typical layout and individual components of a state of the art machine. One or several hydraulic pumps, powered by a diesel engine, provide pressurised flow to the system. One or several hydraulicthe pumps, powered a diesel engine, provide pressurised flow themachine. system. Figure 1 illustrates typical layout by and individual components of a state of thetoart Using joysticks in in thethe cab, thethe operator controls a series of directional valves located in a manifold block, Using cab, operator controls a series of directional valves located a manifold One orjoysticks several hydraulic pumps, powered by a diesel engine, provide pressurised flow in to the system. often referred to as the main control valve (MCV). These allow an intuitive and precise distribution block, often referred asthe theoperator main control valve (MCV). These allow an located intuitiveinand precise of Using joysticks in the to cab, controls a series of directional valves a manifold thedistribution incoming pump flow to the individual actuators of the implement structure (boom, arm, bucket, the incoming flowcontrol to the individual actuators of the implement structure block, oftenofreferred to as pump the main valve (MCV). These allow an intuitive and (boom, precise swing) and travel drive. wheeled machine a bucket attached asimplement a tool is shown figure, arm, bucket, swing) andAtravel drive. A to wheeled machine with a bucket attached asstructure a toolinisthe shown distribution of the incoming pump flow the with individual actuators of the (boom, butin various other arrangements with tracks and other specialised attachments, such as hydraulic thebucket, figure, swing) but various other arrangements withmachine tracks and other specialised attachments, as arm, and travel drive. A wheeled with a bucket attached as a tool issuch shown hammers and scissors, canscissors, be found. hydraulic hammers and can be found.with tracks and other specialised attachments, such as in the figure, but various other arrangements
hydraulic hammers and scissors, can be found.
Figure 1. Layout of a typical hydraulic excavator. Figure 1. Layout of a typical hydraulic excavator. Figure 1. Layout of a typical hydraulic excavator.
The flow of power through the machine and location of the individual losses are depicted in The 2. flow power through the machine location of the individual losses are depicted in Figure Theofconversion of energy one and formlocation into another, performed by the and The flow of power through the from machine and of the as individual losses are engine depicted in Figure 2. The conversion of energy fromThe one form into another, as throttling performedlosses, by theisengine and pump(s), is the first source of losses. referred to as as a direct Figure 2. The conversion of energy from onesecond, form into another, performed by the engine and pump(s), is the first source of losses. The second, referred to as throttling losses, is a direct consequence consequence of using hydraulic power. Finally, thethrottling ancillary drives for pump(s), is the first valves sourcetoofdistribute losses. The second, referred to as losses,required is a direct of vital usingfunctions, valves of tosuch distribute hydraulic power. Finally,power. the required for vital functions, asvalves steering, braking and cooling, are ancillary considered losses as they consume energy consequence using to distribute hydraulic Finally,drives the ancillary drives required for such as steering, braking and cooling, are considered losses as they consume energy and do not directly and do not directly contribute to the execution of the required task. vital functions, such as steering, braking and cooling, are considered losses as they consume energy contribute to the execution of the task.of the required task. and do not directly contribute to required the execution
Figure 2. The flow of power through the machine and the losses along the path. Figure 2. The flow of power through the machine and the losses along the path. Figure 2. The flow of power through the machine and the losses along the path.
Energies 2017, 10, 687
3 of 25
The exact reasons for these losses and their importance can be derived by analysing the individual machine subsystems and their interactions with one another in more detail. Let us begin by discussing the engine. 2.1. Engine In theory, an engine can reach an efficiency of up to 100%, as the chemical input and mechanical output energies are both ordered and of the same quality, i.e., have the same exergy [3,4]. Unfortunately, converting chemical energy directly into mechanical work proves difficult and the fuel must first be ignited in an unconstrained chemical reaction generating heat and losses. As a result, engines specifically designed and optimised to work at a specific operating point can push their peak efficiency into a region of 45–50% but the typical units found in construction machinery display maximum efficiencies of just over 40% [5]. This value changes considerably depending on the load, i.e., torque, and speed at which the engine operates. An exemplary internal combustion engine (ICE) efficiency map is shown in Figure 3a. The map is bounded by the full load line Tmax (n), which corresponds to the maximum torque or power the ICE can generate at each rotation speed. If the ICE is operated along this line, no additional acceleration torque is available. Load torques above this value lead to a deceleration of the output shaft. Only above the minimum rotation speed nmin of about 800 rpm can an engine generate enough mechanical power to overcome its internal losses. Typical units provide a substantial acceleration torque at rotation speeds larger than 1000 rpm [6]. As the speed increases and more power can be generated, the efficiency starts improving. However, higher speeds also lead to increased friction losses and allow less time for the reactants to mix during the combustion process. As a result, efficiency first increases and then begins decreasing with speed. The so-called sweet spot in the contour plot is therefore usually located in the lower to mid-speed range just below the full load line. Efficiency maps are characterised by curves of constant power, hyperbolas. Not every power curve passes through the optimal operating region, indicating that the unit can only function efficiently within a certain power range. When selecting an engine for a specific application, in this case an excavator, the designer must take into account the peak power demand of the hydraulic system driven by the engine. In most engines, the rotation speed nPowMax , at which maximum power is available, is considerably higher than the speed nopt at which maximum efficiency is attained. In order to deliver peak power and thereby avoid having to change the engine speed during a working cycle, standard excavators are frequently operated at a fixed engine speed, namely nPowMax . During such a cycle, the load pressure and pump displacement will vary constantly, leading to fluctuations in the load torque shown in grey in Figure 3a. This results in frequent part loading and therefore inefficient engine utilisation [7]. Thinking in terms of absolute efficiency can be misleading, as fuel consumption is, in fact, the quantity that really matters. The Willans approximation, illustrated in Figure 3b, is another method of plotting engine performance and reveals some interesting aspects that are not so evident from the . contour plotted efficiency map [8]. The fuel consumption V Diesel of an engine at a constant speed increases approximately linearly with its output power. For most diesel engines, this proportionality factor takes on a value of about 0.22 L/kWh [9]. In other words, each additional kW of output power results in the same increase in the fuel consumption rate, regardless of the current operating point. What this actually means is that the engine’s differential efficiency is actually a constant with a value of 42%. The Willans lines also show that an engine delivering no power still consumes fuel due to its . parasitic losses. This is referred to as the idle fuel consumption V 0 and largely depends on the engine size VICE , the number of pistons/cylinders and rotation speed nICE . In general, larger engines and higher rotation speeds lead to higher parasitic losses and therefore an increase in idle fuel consumption.
Energies 2017, 10, 687
4 of 25
Energies 2017, 10, 687
4 of 25
Figure 3. (a) Typical diesel engine efficiency map; and (b) Willans approximation of fuel consumption Figure 3. (a) Typical diesel engine efficiency map; and (b) Willans approximation of fuel consumption adapted from [8]. adapted from [8].
According to Rohde-Brandenburger, in the case of a turbocharged diesel engine, the Willans Accordingcan to Rohde-Brandenburger, approximation be expressed as followsin [8]:the case of a turbocharged diesel engine, the Willans approximation can be expressed as follows [8]: VDiesel l h = V0 l h + 0.22 l kWh · PICE kW (1) . . l l l = V0 + 0.22 · PICE [kW] (1) V Diesel with h h kWh (2) VICE, 0 = VICE · (8·10-8 · nICE 2 + 7.5·10-5 · nICE + 0.1) with . V ICE,0 why = VICE ·(8·with 10−8 ·engine nICE 2 +start–stop 7.5·10−5 ·nsystems ) (2) This is in fact the reason cars fuel consumption. ICE + 0.1reduce Similarly, this explains the real advantage of the concepts referred to as engine downsizing and This is in fact the reason why cars with engine start–stop systems reduce fuel consumption. downspeeding. It is not due to the widely cited claim that the thermodynamic efficiency of the engine Similarly, this explains the real advantage of the concepts referred to as engine downsizing and improves when operating at higher loads and smaller displacements, but rather that the idle fuel downspeeding. It is not due to the widely cited claim that the thermodynamic efficiency of the engine consumption of a downsized or downsped engine is lower [9,10]. This concept reveals invaluable improves when operating at higher loads and smaller displacements, but rather that the idle fuel information regarding how much the fuel consumption of a vehicle can be improved without making consumption of a downsized or downsped engine is lower [9,10]. This concept reveals invaluable changes to the engine. For example, in the case of an average passenger vehicle with a four cylinder information regarding how much the fuel consumption of a vehicle can be improved without making engine, the idle fuel consumption is responsible for up to 50% of the total fuel consumption in the changes to the engine. For example, in the case of an average passenger vehicle with a four cylinder new European driving cycle (NEDC). This would mean that the maximum fuel reduction achievable, engine, the idle fuel consumption is responsible for up to 50% of the total fuel consumption in the by designing a new vehicle with theoretically zero mass and zero wind resistance but without any new European driving cycle (NEDC). This would mean that the maximum fuel reduction achievable, changes to the engine and gearbox, is only 50% [9]. This highlights the importance of using smaller by designing a new vehicle with theoretically zero mass and zero wind resistance but without any engines operating at lower speeds. changes to the engine and gearbox, is only 50% [9]. This highlights the importance of using smaller These observations have important consequences for the energy management strategies in engines operating at lower speeds. hybrid vehicles. When an engine is already running, the decision to charge the battery or accumulator These observations have important consequences for the energy management strategies in hybrid using the engine can be made without regarding the current operating point because the differential vehicles. When an engine is already running, the decision to charge the battery or accumulator using efficiency is constant [11]. Only when the engine is off does the decision to charge lead to additional the engine can be made without regarding the current operating point because the differential efficiency losses due to idling. is constant [11]. Only when the engine is off does the decision to charge lead to additional losses due Reducing engine losses has not been one of the main concerns of excavator manufacturers in to idling. recent years. In the late 20th century, studies proved that the particulate matter (PM), carbon Reducing engine losses has not been one of the main concerns of excavator manufacturers in monoxide (CO) and nitrous oxides (NOX) formed during the combustion of diesel fuel, are all recent years. In the late 20th century, studies proved that the particulate matter (PM), carbon monoxide detrimental to human health [12–14]. In an attempt to improve the air quality in cities and towns, (CO) and nitrous oxides (NOX ) formed during the combustion of diesel fuel, are all detrimental to governments introduced a set of stringent diesel emission standards in the 1990s. The initial TIER 1– human health [12–14]. In an attempt to improve the air quality in cities and towns, governments 3 standards (in Europe Stage I–III), phased in between 1998 and 2008, were met using advanced introduced a set of stringent diesel emission standards in the 1990s. The initial TIER 1–3 standards engine designs and basically no or very simple exhaust gas after-treatment methods. With the (in Europe Stage I–III), phased in between 1998 and 2008, were met using advanced engine designs introduction of the TIER 4 interim and final (European Stage IIIB and IV) standards, which were and basically no or very simple exhaust gas after-treatment methods. With the introduction of the phased in from 2008 to 2015, manufacturers first agreed to decrease PM and then NOX emissions by TIER 4 interim and final (European Stage IIIB and IV) standards, which were phased in from 2008 to a further 90%. Meeting these standards was considerably more difficult. Today’s excavators and other off-highway machines use expensive after-treatment systems with diesel oxidation catalysts (DOC) and diesel particulate filters (DPF). This has led to increased costs
Energies 2017, 10, 687
5 of 25
and a further reduction in the available installation space around the engine. As explained by Filla, 2015, manufacturers first agreed PM and then has NOalso by a further 90%. Meeting X emissions the way in which emissions tests to aredecrease currently conducted given OEMs (Original Equipment these standards was considerably more difficult. Manufacturers) little or no incentive to introduce efficient downspeed engines with hybrid Today’s [15]. excavators and other off-highway use expensive after-treatment systems technologies The official non-road transient machines cycle (NRTC), used to evaluate engine emissions, with diesel oxidation catalysts (DOC) and diesel particulate filters (DPF). This has led to increased uses a predetermined set of engine speeds and torques, which may not even match the actual engine costs and in a further reduction in thea available installation spaceengine aroundmanagement the engine. system As explained operation a machine. As a result, machine with an advanced would by Filla, the way in which emissions tests are currently conducted has also given OEMs (Original be no different to a standard unit in regard to the current engine emissions regulations. Equipment littleabove or no relate incentive to introduce efficient downspeedAs engines hybrid All theManufacturers) aspects mentioned to quasi-static engine performance. shownwith in Figure 4 technologies [15]. Theengine officialloading non-road transient cyclecannot (NRTC), used to evaluate the effect of dynamic should not and be neglected. In any engine typical emissions, excavator uses acycle, predetermined engine speeds and torques, which maybenot even match the actualtoengine duty the engine set loadoffluctuates rapidly and the engine must able to respond quickly these operation in a machine. As a result, a machine with an advanced engine management system would load changes to avoid stalling or excessive drops in rotation speed [16]. be noDepending different toon a standard in regard to the current engine emissions regulations. the extentunit of these transients, fuel consumption can fluctuate considerably from All the aspects mentioned above relate to quasi-static engine performance. shown in Figure the quasi-static descriptions in Figure 3. Research has, in fact, shown that up to As 50% of emissions are4 the effect of dynamic engine loading should not and cannot be neglected. In any typical excavator caused by transient loads [17,18]. As a result, newer TIER 4 engines do not have the same response duty the cycle, the engine fluctuates the engine mustatbelower able toengine respond quickly to these that older TIER 3load engines had. rapidly This is and especially evident speeds and is an load changes to avoid stalling or excessive drops in rotation speed [16]. important issue that will affect the design of newer mobile hydraulic systems.
Figure during aa typical typical excavator excavator cycle. cycle. Figure 4. 4. Measured Measured engine engine load load and and rotation rotation speed speed during
The mechanical power generated by the engine, must now be distributed to the individual Depending on the extent consumption hydraulic actuators. This takesof usthese to thetransients, machine’sfuel hydraulic system.can fluctuate considerably from the quasi-static descriptions in Figure 3. Research has, in fact, shown that up to 50% of emissions are caused by transient loads [17,18]. As a result, newer TIER 4 engines do not have the same response that 2.2. Hydraulic System the older TIER 3 engines had. This is especially evident at lower engine speeds and is an important in which the hydraulic interact systems. with the external environmental load is issueThe that way will affect the design of neweractuators mobile hydraulic particularly complex. Not only are the force and velocity of each actuator The mechanical power generated by the engine, mustdemands now be distributed to the completely individual different, they also vary independently of each other depending on the operator’s commands. Some hydraulic actuators. This takes us to the machine’s hydraulic system. actuators may require high force and low velocity (high pressure, low flow) while others require low force and high velocity (low pressure, high flow). Figure 5 illustrates the load situation for the 2.2. Hydraulic System actuators making up the implement structure. The x-axis shows the flow required by each actuator The way in which the hydraulic actuators interact with the external environmental load is (QL ) and can be interpreted as the operator’s input to the system. The y-axis shows the force or load particularly complex. Not only are the force and velocity demands of each actuator completely pressure (pL ) experienced by the actuator and is a direct consequence of the surroundings, cf. different, they also vary independently of each other depending on the operator’s commands. Some Equations (3)–(5). actuators may require high force and low velocity (high pressure, low flow) while others require low ARod force and high velocity (low pressure, flow). the load situation for the actuators pLhigh = pPiston - Figure·5pillustrates (3) APiston Rod making up the implement structure. The x-axis shows the flow required by each actuator (QL ) and can be interpreted as the operator’s input to the system. The y-axis shows the force or load pressure (pL ) QL = QPiston (4) experienced by the actuator and is a direct consequence of the surroundings, cf. Equations (3)–(5). PAct = FL ·x = pL APiston ·x = pL QPiston = pL QL (5) ARod pPiston − acting on· pthem pL = the (3) In the case of the linear actuators, force APiston Rod is due to both the weight of the attached structure and the external forces present during digging and other operations. Inertial forces caused during acceleration play a less important role. Depending on the movements, each actuator
Energies 2017, 10, 687
6 of 25
QL = QPiston .
.
PAct = FL · x = pL APiston · x = pL QPiston = pL QL
(4) (5)
In the case of the linear actuators, the force acting on them is due to both the weight of the attached structure and the external forces present during digging and other operations. Inertial forces Energies 2017, 10, 687 6 of 25 caused during acceleration play a less important role. Depending on the movements, each actuator experiences experiences either either aa resistive resistive force force opposing opposing its its motion motion (Quadrants (Quadrants II and and III) III) or or an an assistive assistive force force aiding II and aiding its its motion motion (Quadrants (Quadrants II and IV). IV). Consequently, Consequently, in in quadrants quadrants II and and III, III, the the actuator actuator must must be be actively supplied with power, while, in quadrants II and IV, the actuators can actually supply power to actively supplied with power, while, in quadrants II and IV, the actuators can actually supply power the system. to the system.
Figure 5. Load quadrants experienced by actuators of the implement system. Figure 5. Load quadrants experienced by actuators of the implement system.
Due to the kinematic arrangement and large weight of the implement structure, the boom Due to the kinematic arrangement and large weight of the the boom cylinders almost exclusively operate in load quadrants I and II. implement In contrast, structure, the magnitude and cylinders almost exclusively operate in load quadrants I and II. In contrast, the magnitude and direction of the load acting on the arm and bucket cylinders varies a great deal, causing operation in direction of the loadThe acting on the motor arm and bucket cylinders varies a great deal, causing operationbut in all four quadrants. hydraulic driving the swing experiences four quadrant operation, all four quadrants. The hydraulic motor driving the swing experiences four quadrant operation, but in contrast to the linear actuators the inertial forces dominate here, meaning that the load pressure in is contrast to the linear actuators the inertial forces dominate here, meaning that the load pressure is mainly due to the acceleration of the superstructure and not caused by external forces. mainly duepoint to theinacceleration of therepresents superstructure and caused by external forces. in which the Each the pL /QL plane a state ofnot quasi-stationary equilibrium, Each point in the pL /QL plane represents a state of quasi-stationary equilibrium, in which the pump flow rate is proportional to the operator’s joystick displacement, system pressure is determined pump flow rate is proportional to the operator’s joystick displacement, system pressure is determined by the load, and the engine torque and pump torque are equal. As the actuator operating points move by the load, and the engine torque and pump torque are equal. As the actuator operating points move through the different load quadrants, the system’s power demand changes. To maintain a stable through the different load quadrants, the system’s power demand changes. To maintain a stable engine engine speed, every change in demand must be closely followed by a change in supply. This can be speed, every change in demand must be closely followed by a change in supply. This can be tricky for tricky for two reasons. two reasons. Firstly, the engines in most machines cannot even deliver the same amount of power as can be Firstly, the engines in most machines cannot even deliver the same amount of power as can be demanded by the pump. This may sound strange, but has to do with dimensioning. The pump size demanded by the pump. This may sound strange, but has to do with dimensioning. The pump size is is selected to meet the maximum speed/flow rate Qmax requirements of all the actuators selected to meet the maximum speed/flow rate ( Qmax ) requirements of all the actuators (implements (implements + travel drive) when operating at the rated engine speed. System pressure is a function + travel drive) when operating at the rated engine speed. System pressure is a function of the load of the load and can reach values up to pmax = 380 bar, but is typically around 200 bar. The pump’s and can reach values up to pmax = 380 bar, but is typically around 200 bar. The pump’s corner power corner power is, therefore, considerably greater than the power required during standard operation. is, therefore, considerably greater than the power required during standard operation. Installing an Installing an engine with the same corner power capabilities as the pump in order to cover these engine with the same corner power capabilities as the pump in order to cover these infrequent cases infrequent cases would expensive and require space. a result, it is quite common to find would be expensive andberequire more space. Asmore a result, it isAsquite common to find machines, in machines, in which the corner power of the pump is two or three times greater than the maximum which the corner power of the pump is two or three times greater than the maximum engine power. engine power. Full actuator speed can only be maintained for average system pressures around 180 bar, thereafter a power limiting controller swivels the pump back preventing the engine from stalling. Secondly, as the operator adjusts his joystick commands and the load changes, an actuator’s state can move through the load plane rather rapidly. Unfortunately, most engines cannot react so quickly to such rapid load changes and are prone to stalling. The pump controller and valves must assume
Energies 2017, 10, 687
7 of 25
Full actuator speed can only be maintained for average system pressures around 180 bar, thereafter a power limiting controller swivels the pump back preventing the engine from stalling. Secondly, as the operator adjusts his joystick commands and the load changes, an actuator’s state can move through the load plane rather rapidly. Unfortunately, most engines cannot react so quickly to such rapid load changes and are prone to stalling. The pump controller and valves must assume this role, and ideally work in sync with the EECU, to ensure the dynamic power demand and supply are well matched. Achieving a well-tuned engine-pump interface is one of the major challenges in Energies 2017, 10, 687 7 of 25 today’s mobile hydraulic systems [19]. In most otherother mobilemobile machines, excavators are capableare of rotating In contrast contrastto to most machines, excavators capable their of superstructure rotating their relative to the undercarriage by using the swing drive. The way in which the acceleration superstructure relative to the undercarriage by using the swing drive. The way in which and the deceleration of the swing is regulated largely determines the machine’s performance. Not only is this acceleration and deceleration of the swing is regulated largely determines the machine’s drive used approximately 60%drive of theused time,approximately it is also critical to of safety because operator’s of performance. Not only is this 60% the time, it isthe also critical tofield safety vision changes as the machine turns. As a result, swing motion must be precise and have priority because the operator’s field of vision changes as the machine turns. As a result, swing motion must over other and functions. Each OEM itsfunctions. own specific solution, to its attain required swingtomotion be precise have priority over has other Each OEM has ownthe specific solution, attain without affecting the other actuators negatively. The single circuit flow sharing (SC-FS) system studied the required swing motion without affecting the other actuators negatively. The single circuit flow in this work, which also represents thethis mostwork, commonly system wheeled excavators, sharing (SC-FS) system studied in whichused alsohydraulic represents the for most commonly used is shown insystem Figurefor 6. wheeled excavators, is shown in Figure 6. hydraulic
Figure 6. Single Circuit Flow Sharing (SC-FS) circuit used in European wheeled excavators, including Figure 6. Single Circuit Flow Sharing (SC-FS) circuit used in European wheeled excavators, including downstream pressurecompensators compensators (PCs) load independent control, a torque controlled (TC) downstream pressure (PCs) forfor load independent control, a torque controlled (TC) swing B). swing with upstream compensation and LS-pressure bypass on the boom (LS with upstream compensation and LS-pressure bypass on the boom (LSB ).
The implement cylinders are controlled using downstream stream pressure compensated The implement cylinders are controlled using downstream stream pressure compensated valves. valves. Controlling the swing’s large inertia with such a flow controlled valve, causes high pressures Controlling the swing’s large inertia with such a flow controlled valve, causes high pressures during during acceleration as the load pressure is primarily determined by inertial forces resulting from acceleration as the load pressure is primarily determined by inertial forces resulting from acceleration acceleration and not by external load forces from the environment [20]. If the acceleration is not and not by external load forces from the environment [20]. If the acceleration is not controlled controlled appropriately, the system pressure can rise excessively leading to large pressure appropriately, the system pressure can rise excessively leading to large pressure differences among differences among the various actuators, which cause unnecessary throttling and forces the pump to the various actuators, which cause unnecessary throttling and forces the pump to swivel back thereby swivel back thereby delivering less flow in order to avoid overloading the engine [21]. Such issues delivering less flow in order to avoid overloading the engine [21]. Such issues are detrimental to are detrimental to performance. To overcome these issues, manufacturers have come up with various performance. To overcome these issues, manufacturers have come up with various solutions. solutions. Instead regulating the to the the swing, swing, so-called so-called torque torque control control is is used. used. In In these these circuits, circuits, Instead of of regulating the flow flow rate rate to the operator’s joystick command determines the maximum load sensing pressure of the swing the operator’s joystick command determines the maximum load sensing pressure of the swing drive drive and is used to directly regulate the swing torque and not the swing speed. To ensure the swing and is used to directly regulate the swing torque and not the swing speed. To ensure the swing always always the it flow it demands, regardless of the other actuatormovements, movements,an an upstream upstream receivesreceives exactlyexactly the flow demands, regardless of the other actuator pressure compensator with a lower spring pressure differential setting than the pump controller’s pressure compensator with a lower spring pressure differential setting than the pump controller’s ∆p is used used [22]. [22]. This This gives gives the the swing swing priority priority over over the the other This is is an an important LS is ∆pLS other actuators. actuators. This important safety safety function as the operator’s field of view changes during swing operation and unexpected obstacles may suddenly appear. An additional feature of the circuit is the boom load sensing pressure bypass. During fast lifting operations, the load sensing line is directly connected to the boom piston side using a bypass throttle,
Energies 2017, 10, 687
8 of 25
function as the operator’s field of view changes during swing operation and unexpected obstacles may suddenly appear. An2017, additional Energies 10, 687 feature of the circuit is the boom load sensing pressure bypass. During fast lifting 8 of 25 operations, the load sensing line is directly connected to the boom piston side using a bypass throttle, To the discuss losses occurring within thepressure hydraulic system, wesent must takepump. a closer look at the causing boomthe pressure to be the dominant signal that is to the This ensures pumps, valvespressure and actuators. that the swing does not exceed the boom pressure, which results in minimal throttling and maximum pump flow, improving cycle times and efficiency [20]. Some manufacturers also offer an 2.2.1. Pump Losses energy efficient boom float function in which the boom down motion does not require any pump flow [23]. These features represent state of the art in today’s wheeled The hydraulic pump, almost the exclusively of the axial piston type, isexcavators. responsible for converting To discuss the losses occurring within the hydraulic system, we takeofa flow closer look at the the mechanical output power from the engine into hydraulic power, inmust the form and pressure. pumps, valves and actuators. A typical pump efficiency map for a constant rotation speed is shown in Figure 7a. Depending on the
Displacement Set ting [-]
displacement setting and pressure, the efficiency can vary from as low as 60% to peak values of up 2.2.1. Pump Losses to 91% at higher displacement settings and pressure levels [24]. As with the engine, thinking in terms The hydraulic almost exclusively of the axial piston type, is responsible for converting of efficiency can bepump, misleading. The leakage and hydro-mechanical losses do not change depending the outputsetting, power from the engine into hydraulic power,In inreality, the form of flow and pressure. on mechanical the displacement they actually remain fairly constant. only the output power Achanges typical leading pump efficiency map for a constant rotation speed showndisplacement in Figure 7a. Depending on the to a higher efficiency. A pump operating at is a higher actually consumes displacement setting and pressure, efficiency can vary from as low as 60% to peak values of up to more energy as its power output isthe higher. 91% at higher displacement levels Asalso withcreate the engine, terms The mismatched cornersettings powersand of pressure the engine and [24]. pump some thinking additionalineffects. of efficiency can be misleading. The leakage and hydro-mechanical losses do not change depending Pump operation in the upper right hand region is not even possible, meaning that the pump is forced on the displacement setting, theysettings actuallyand remain fairly constant. In reality, only output power to operate at lower displacement efficiencies when system pressure is the high. Studies have changes leading to a higher efficiency. A pump at amobile higherhydraulic displacement actually consumes shown that depending on the cycle the pumpsoperating in a typical system are responsible more energy as between its power10% output higher. for dissipating andis15% of the mechanical power supplied by the engine [25,26]. (a)
1
(b) PPump,max
0,91 0,90
PICE,max
Displacement act uator and cont roller pLS
0,89 0,87
0,5
0,85 0,8
0
0,75 0,7
Standard Operating Region
180
360 Pressure [bar]
0 ... 1
Damping Orifice
Act ual pump
Figure7.7.(a) (a)Characteristic Characteristic efficiency efficiency map map for for aa200 200cc ccaxial axialpiston piston unit unit[24]; [24];and and(b) (b)aaschematic schematicof ofaa Figure pump pumpwith withaaload loadsensing sensingpressure pressurecontroller. controller.
In the past couple of years, oneoffurther aspect has become The mismatched corner powers the engine andconcerning pump alsopump create operation some additional effects.known. Pump The values shown in typical pump efficiency maps are taken from measurements, in which pump operation in the upper right hand region is not even possible, meaning that the pump is the forced to controller is inactive and the displacement actuator has been mechanically locked thereby not operate at lower displacement settings and efficiencies when system pressure is high. Studies have allowing pump swash vibrate. are notmobile realistic boundary conditions and do for not shown thatthe depending on theplate cycletothe pumpsThese in a typical hydraulic system are responsible represent how such units really operate in a machine. The pump’s controller is, in fact, constantly dissipating between 10% and 15% of the mechanical power supplied by the engine [25,26]. adjusting swash plate and regulating theaspect flow entering the system. In thethe past couple of years, one further concerning pump operation has become known. This causes two additional loss mechanisms. First, the controller has The values shown in typical pump efficiency maps are taken hydro-mechanical from measurements, in whichusually the pump various damping orifices (Figure 7b), which createhas additional leakage. The second loss mechanism controller is inactive and the displacement actuator been mechanically locked thereby not allowingis due to the dynamic high frequency oscillation of the swash plate [27]. As the pump piston barrel the pump swash plate to vibrate. These are not realistic boundary conditions and do not represent rotates, a strong vibrating torque acts on the swash plate. In the case of tests, in which the swash how such units really operate in a machine. The pump’s controller is, in fact, constantly adjustingplate the is mechanically locked, this vibrating torque is counteracted by pump’s end stop. In reality, the pump swash plate and regulating the flow entering the system. controller must feed the swash plate with pressure in order to balance these torque oscillations, thereby increasing the energy consumed by the controller. Together these controller losses are by no means negligible and can decrease efficiency by up to ten percentage points [27,28]. Just like in a diesel engine, the concept of efficiency can be a bit misleading. A variable displacement pump operating in standby at near 0 displacement still consumes power. For example,
Energies 2017, 10, 687
9 of 25
This causes two additional loss mechanisms. First, the hydro-mechanical controller usually has various damping orifices (Figure 7b), which create additional leakage. The second loss mechanism is due to the dynamic high frequency oscillation of the swash plate [27]. As the pump piston barrel rotates, a strong vibrating torque acts on the swash plate. In the case of tests, in which the swash plate is mechanically locked, this vibrating torque is counteracted by pump’s end stop. In reality, the pump controller must feed the swash plate with pressure in order to balance these torque oscillations, thereby increasing the energy consumed by the controller. Together these controller losses are by no means negligible and can decrease efficiency by up to ten percentage points [27,28]. Just like in a diesel engine, the concept of efficiency can be a bit misleading. A variable displacement pump operating in standby at near 0 displacement still consumes power. For example, a 210 cc pump operating at 1800 rpm and a standby pressure of 28 bar will consume around 4 kW of power. These parasitic losses increase with rotation speed and cannot be neglected as they increase the machine’s idle fuel consumption. A Willans representation of pump efficiency has yet to published, but would surely be extremely valuable. 2.2.2. Valve Losses The flow leaving the pump(s) is distributed to the actuators using valves. The fundamental physics explaining the flow of fluid through a valve can be described using the orifice equation: Q = KV y
p
∆p
(6)
In order for a flow Q to pass through a valve a certain pressure difference ∆p, in other words a driving force, must be present. The amount of pressure needed depends on the valve’s geometry, described by the coefficient KV , and spool position y. In summary, for a valve to function, part of the hydraulic power entering the valve must be dissipated as heat. These so-called pressure or throttling losses can be expressed as follows PThrottle = Q∆p (7) In literature, there seems to be a number of misconceptions concerning throttling. Yes, due to their hydraulic resistance valves will always generate losses when supplying flow to an actuator, but if correctly sized, that is with a large enough value for KV , these losses can be kept low, down to only a couple of bar. The extreme throttling losses attributed to valves and so often cited in literature, are due to completely different reasons. These are worth mentioning. The first and major cause is directly related to the nature of the hydraulic architecture. Multiple actuator systems, in which valves are used to distribute flow, delivered by a single pump are a classic example. Each actuator has its own independent pressure level, determined by the load it is currently subjected to. In order to deliver flow to each actuator the pump must supply the system with a pressure level greater than that of each actuator. Unfortunately, no matter how well the cylinder areas and kinematics are designed, there will always be instances, during which the individual actuator pressure levels are substantially different. This creates a complete mismatch between the pump pressure and actuator pressures. Unfortunately, valves can only operate according to the orifice equation and are not capable of performing a lossless transformation of the pump pressure down to a lower actuator pressure level. Consequently, the already existing pressure difference is used to generate the required flow, resulting in considerable losses, especially when the flow is large, see Equation (7). Other than distributing flow, the valves must enable the operator to precisely regulate the movement of an actuator in all four load quadrants. This is done by regulating both the flow to the actuator using an inlet metering edge and the flow leaving the actuator using an outlet metering edge. The inlet edge is needed to control the actuator speed as it determines the amount of flow entering the actuator. The outlet edge, on the other hand, fulfils a different function. In quadrants II and IV, it is used as a type of brake and prevents runaway loads. In quadrants I and III, it is used to marginally throttle the flow leaving the actuator, thereby maintaining a pressure level of
Energies 2017, 10, 687
10 of 25
approximately 5 to 10 bar in the outlet chamber [29]. This ensures that both the inlet and outlet chambers are always pressurized and act as springs, guaranteeing a higher natural frequency and better system response [30]. For actuators undergoing four quadrant operation, finding an optimal Energies 2017, 10, 687 involves compromise. The outlet resistance must be large enough to prevent 10 ofthe 25 outlet geometry overrunning loads in quadrants II and IV, but preferably should not be too large to cause unnecessary 2.3. Ancillary Drives throttling in the other two quadrants I and III. In summary, it is hydraulic important system, to differentiate between following causes Apart from the the engine alsothe provides power to of a throttling: number of smaller
subsystems, referred asinlet ancillary basically only consume energy and do not directly • Throttling acrosstothe edgedrives, in orderwhich to supply the actuator with flow contribute in the generation of useful work, but without them the machine could not function. A • Throttling across the outlet edge to maintain controllability and prevent runaway loads typical setup is shown schematically in Figure 8. • Throttling to equalise mismatched supply and actuator pressures These include: Drivesregulating the machine’s 24 V electrical system; 2.3. Ancillary The alternator The airfrom conditioning compressor unit; Apart the hydraulic system, the engine also provides power to a number of smaller subsystems, referred A number of hydraulic gear pumps providing pressure for and the joysticks as well as for the to as ancillary drives, which basically only consume energy do not directly contribute in steering, braking and cooling systems. It should be noted that crawler type excavators do not the generation of useful work, but without them the machine could not function. A typical setup is have a separateinsteering shown schematically Figure 8.system and therefore do not require a pump supplying steering pressure.
Figure 8. Ancillary drives in a wheeled excavator. Figure 8. Ancillary drives in a wheeled excavator.
Figure 9 shows measurements of engine output power in an excavator at idle for different engine include: speedThese settings. To keep all the components on the shaft, including the ancillary drives and main pump, rotating up to 30% of the full24 output powersystem; is consumed. • The alternator regulating theengine’s machine’s V electrical
• •
The air conditioning compressor unit; A number of hydraulic gear pumps providing pressure for the joysticks as well as for the steering, braking and cooling systems. It should be noted that crawler type excavators do not have a separate steering system and therefore do not require a pump supplying steering pressure.
Figure 9 shows measurements of engine output power in an excavator at idle for different engine speed settings. To keep all the components on the shaft, including the ancillary drives and main pump, rotating up to 30% of the engine’s full output power is consumed. In mathematical terms, the idle power demand can be described as a linear function of the engine speed: PIdle = m0 ·speed nICE during a typical excavator cycle. (8) Figure 9. Measured engine load and rotation In order to ensure no damage occurs if the machine is operated at low speeds, the ancillary drives In mathematical terms, the idle power demand can be described as a linear function of the engine are dimensioned to be fully functional at lower engine speeds of 800 rpm. Therefore, the additional speed: ancillary power produced at higher engine speeds is totally unnecessary and can be considered as an additional loss term. (8) PIdle = m0 · nICE
In order to ensure no damage occurs if the machine is operated at low speeds, the ancillary drives are dimensioned to be fully functional at lower engine speeds of 800 rpm. Therefore, the additional ancillary power produced at higher engine speeds is totally unnecessary and can be considered as an additional loss term.
Figure 8. Ancillary drives in a wheeled excavator.
Figure 9 shows measurements of engine output power in an excavator at idle for different engine Energies 2017, settings. 10, 687 To keep all the components on the shaft, including the ancillary drives and main 11 of 25 speed
pump, rotating up to 30% of the engine’s full output power is consumed.
Figure 9. Measured engine load and rotation speed during a typical excavator cycle.
Figure 9. Measured engine load and rotation speed during a typical excavator cycle.
In mathematical terms, the idle power demand can be described as a linear function of the engine
3. Cycle Analysis speed: Energies 2017, 10, 687
11 of 25
PIdleof =m Hydraulic excavators, especially in the case those up to 25 tons, are used for a variety of (8) different 0 · nICE tasks3.not only for digging and moving earth. As a result, engineers have to deal with the complex Cycle Analysis In order to ensure no damage occurs if the machine is operated at low speeds, the ancillary drives task of to design machine without knowledge ofup itsof exact future application. Fortunately, arehaving dimensioned to beafully functional at the lower engine speeds rpm. Therefore, thea additional Hydraulic excavators, especially in case of those to 800 25 tons, are used for variety of collecting detailed information on fuel consumption and important states, is becoming ancillary power produced higher engine speedsearth. is totally and system can be considered asthe an different tasks not only foratdigging and moving Asother aunnecessary result, engineers have to deal with loss term. easieradditional with the introduction of fleet management systems. These allow companies to track and record complex task of having to design a machine without knowledge of its exact future application. the movements well as detailed operation of all their [31]. Without a doubt, thissystem new technology Fortunately,as collecting information on machines fuel consumption and other important states, is becoming with the introduction of fleetasmanagement systems. allow to will help change easier the way machines are designed, engineers learn howThese to mine andcompanies extract valuable track and record themassive movements well as operation all their machines [31]. Without a doubt,results this for knowledge from these dataasquantities. In 2013,ofLiebherr published some interesting new technology help change the way machines designed, as engineers learn how to spent mine 25% a wheeled excavator will obtained using their LiDat systemare [32]. These showed that the machine and extract valuable knowledge from these massive data quantities. In 2013, Liebherr published some of its operating time idling, 15% travelling between worksites and the remaining 60% performing actual interesting results for a wheeled excavator obtained using their LiDat system [32]. These showed that earth moving tasks. Using a similar fleet management system, researchers in Wuppertal collected the machine spent 25% of its operating time idling, 15% travelling between worksites and the data for 3733 machines over a period of sixmoving monthstasks. [33]. Using For wheeled they found that, on remaining 60% performing actual earth a similarexcavators, fleet management system, average, these machines spent 30% of their operating time idling, 20% travelling, 10% grading researchers in Wuppertal collected data for 3733 machines over a period of six months [33]. For and 40% digging. Tracked excavators 15% at idle, 10%machines travelling, 15% grading the remaining wheeled excavators, they foundspent that, on average, these spent 30% of their and operating time 60% of their20% operating time10% digging. idling, travelling, grading and 40% digging. Tracked excavators spent 15% at idle, 10% travelling, 15% grading and the remaining 60% of their operating time digging.
3.1. Characteristics of Typical Duty Cycles 3.1. Characteristics of Typical Duty Cycles
Although not exhaustive, a survey of research conducted in this field suggests that dig and Althoughand notgrading exhaustive, a survey research conducted in thisperformed field suggests that dig and dump, trenching, are the three of most common duty cycles by excavators [21,33]. dump, trenching, and grading are the three most common duty cycles performed by excavators These are illustrated in Figure 10. An analysis of measurement data, collected from an 18 t excavator [21,33]. These are illustrated in Figure 10. An analysis of measurement data, collected from an 18 t performing these three cycles, shows some typical aspects of excavator operation, which can help excavator performing these three cycles, shows some typical aspects of excavator operation, which explain the main cause of losses and identify potential for improvement. can help explain the main cause of losses and identify potential for improvement.
Figure 10. Typical excavator duty cycle.
Figure 10. Typical excavator duty cycle.
The first important characteristic to take note of is the large fluctuation in actuator power demand for all three cycles. Tableto1take summarises fluctuations usinginthree parameters: zP The first important characteristic note of isthese the large fluctuation actuator power demand the ratio of the1 mean and maximum actuator power demand during the cycle, zzPQ describes for alldescribes three cycles. Table summarises these fluctuations using three parameters: describes the the ratio of the mean and maximum pump flow rate during the cycle and finally zp describes the ratio of the mean and maximum pump pressure during the cycle. The peak power required for each of the cycles is at least five greater than the mean power output. Similarly, the peak values for the pump flow rate and pressure are approximately two times greater than the mean values. Table 1. Comparison of average and maximum power, flow and pressure for each cycle.
Energies 2017, 10, 687
12 of 25
ratio of the mean and maximum actuator power demand during the cycle, zQ describes the ratio of the mean and maximum pump flow rate during the cycle and finally zp describes the ratio of the mean and maximum pump pressure during the cycle. The peak power required for each of the cycles is at least five greater than the mean power output. Similarly, the peak values for the pump flow rate and pressure are approximately two times greater than the mean values. Table 1. Comparison of average and maximum power, flow and pressure for each cycle. Cycle
zP [−]
zQ [−]
zp [−]
Dig & Dump Trenching Grading
0.20 0.18 0.08
0.60 0.50 0.45
0.51 0.50 0.44
These observations should come as no surprise because a machine like an excavator can, in fact, Energies 2017, 10, 687 12 of 25 only perform cyclic tasks. Every time the boom is raised, it must eventually be lowered. Similarly, every acceleration of the swing be followed its deceleration. Everycan, period of high These observations shoulddrive come must as no surprise becauseby a machine like an excavator in fact, poweronly demand is followed by a period in which energy is no longer required or can actually be perform cyclic tasks. Every time the boom is raised, it must eventually be lowered. Similarly, everyfrom acceleration of the swing drive exception must be followed its isdeceleration. periodduring of highwhich recovered the actuator. The only to this by rule travelling Every operation, power demand is followed a period in whichofenergy is noalllonger or can actually be will constant power is required for aby prolonged period time. For otherrequired cycles, the power demand recovered from the actuator. The only exception to this rule is travelling operation, during which fluctuate considerably. constant power is required for a prolonged period of time. For all other cycles, the power demand An important consequence of the fluctuating power is that all the components in the system must will fluctuate considerably. be designed peak requirements but spend most time operating at part load. Anfor important consequence of thewill fluctuating powerofistheir that all theactually components in the system As will be discussed on in this section, is associated lowoperating efficiency for must be designedlater for peak requirements butpart will loading spend most of their timewith actually at values part both the engine pump. later on in this section, part loading is associated with low efficiency values load. As willand be discussed for both the engine and typical pump. for multi-actuator hydraulic systems, is illustrated for each cycle The second property, The second property, typical for kinematic multi-actuator hydraulic systems, illustratedarm for each and actuator in Figure 11. Due to the arrangement of the isexcavator the cycle load force and actuator in Figure 11. Due to the kinematic arrangement of the excavator arm the load force acting on an actuator not only varies during a duty cycle but is also different compared to the load acting on an actuator not only varies during a duty cycle but is also different compared to the load force experienced by the other actuators. This means that each actuator has its own unique power force experienced by the other actuators. This means that each actuator has its own unique power demand profile, which can be expressed in terms of load pressure and required flow rate ( p , QL ). demand profile, which can be expressed in terms of load pressure and required flow rate pL , QL L. As discussed previously, using valves to distribute flow from actuators with As discussed previously, using valves to distribute flow froma asingle singlepump pumpto to multiple multiple actuators mismatched power profiles inefficient. with mismatched power is profiles is inefficient.
Figure 11. Load pressure over flow required for each actuator and cycle.
Figure 11. Load pressure over flow required for each actuator and cycle. Before continuing, an important aspect must be mentioned. Yes, for the majority of applications power demand does fluctuate considerably and the average engine output power is lower than its rated power; nevertheless, a machine must be capable of supplying the rated power continuously for prolonged periods of time. For example, when travelling at full speed (30–35 km/h) in a wheeled excavator or for tasks requiring the use of a special attachment like a drill, during which a continuously high pump flow rate is demanded. This must be considered when designing any new
Energies 2017, 10, 687
13 of 25
Before continuing, an important aspect must be mentioned. Yes, for the majority of applications power demand does fluctuate considerably and the average engine output power is lower than its rated power; nevertheless, a machine must be capable of supplying the rated power continuously for prolonged periods of time. For example, when travelling at full speed (30–35 km/h) in a wheeled excavator or for tasks requiring the use of a special attachment like a drill, during which a continuously high pump flow rate is demanded. This must be considered when designing any new system. 3.2. Energy Recovery in Excavators The energy available for recovery in an excavator is found in two forms, either gravitational Energies 2017, 10, 687 or mechanical kinetic energy. When lowering a body of mass m by a distance 13 of∆h 25 potential energy through a gravitational field of strength g the energy Epot , described by Equation (9), is released. through gravitational field mass of strength the energy Epot ,J,described is released. Similarly,a when a body, with m andgrotational inertia traveling by at aEquation speed v, (9), or rotating at a Similarly, when a body, with mass m and traveling at a speed v, or rotating at a speed ω, decelerates its kinetic energy Ekin rotational (Equation inertia (10)) isJ,made available. speed ω, decelerates its kinetic energy Ekin (Equation (10)) is made available. Epot = mg∆h (9) Epot = mg∆h (9) 1 12 11 22 Jω (10) Ekin, lin (10) Ekin,=lin 2=mvmv, 2E,kin,rot Ekin,rot== 2Jω 2 2 It is is worth worth taking taking aa moment moment to to appreciate appreciate the the similarities similarities and and differences differences in in these these two two equations. equations. It Both are linearly dependent on the mass or inertia. Similarly, a change in height affects the Both are linearly dependent on the mass or inertia. Similarly, a change in height affects the potential potential energy linearly, but a change in speed affects the kinetic energy quadratically. Consequently, changes energy linearly, but a change in speed affects the kinetic energy quadratically. Consequently, changes in mass or height have less of an influence compared to changes in speed. in mass or height have less of an influence compared to changes in speed. Due to to its mass and and large the boom Due its substantial substantial mass large changes changes in in height, height, the boom structure structure has has the the greatest greatest amount of recoverable potential energy, Figure 12a. On the other hand, due to the low speed of the the amount of recoverable potential energy, Figure 12a. On the other hand, due to the low speed of centre of mass its kinetic energy is negligible. Similarly, potential energy can be recovered from the centre of mass its kinetic energy is negligible. Similarly, potential energy can be recovered from the arm and toto thethe boom. The swing drives possesses no arm and bucket bucket actuator, actuator,but butconsiderably considerablyless lesscompared compared boom. The swing drives possesses potential energy, but due to its large rotational inertia J, a considerable amount of kinetic energy is no potential energy, but due to its large rotational inertia J, a considerable amount of kinetic energy generated as the superstructure is accelerated (Figure 12b). is generated as the superstructure is accelerated (Figure 12b).
Figure 12. (a) Boom potential energy; and (b) swing kinetic energy. Figure 12. (a) Boom potential energy; and (b) swing kinetic energy.
If the energy released during lowering and braking cannot be reused immediately or stored, it energy released during lowering and braking cannot reusedvalve immediately or machine, stored, it mustIfbethe dissipated and leave the system in the form of heat. In a be standard controlled must be dissipated andisleave system into in the form of heat. In in a standard controlled machine, the recoverable energy first the converted hydraulic power, the form valve of flow Q and pressure p, the recoverable energy is first converted into hydraulic power, in the form of flow Q and pressure p, which is then throttled to tank pressure thereby generating heat and causing an unnecessary increase which is then throttled to tank in oil temperature (Figure 13). pressure thereby generating heat and causing an unnecessary increase in oil temperature (Figure 13).
If the energy released during lowering and braking cannot be reused immediately or stored, it must be dissipated and leave the system in the form of heat. In a standard valve controlled machine, the recoverable energy is first converted into hydraulic power, in the form of flow Q and pressure p, Energies 2017, 10, throttled 687 14 of 25 which is then to tank pressure thereby generating heat and causing an unnecessary increase in oil temperature (Figure 13).
Figure 13. (a) Energy dissipation during: boom lowering; and (b) swing braking. Figure 13. (a) Energy dissipation during: boom lowering; and (b) swing braking.
Numerous papers and patents discussing ways to recover boom and swing energy have been published in recent years [34–39]. Unfortunately, few of these publications answer the following four fundamental questions: 1. 2. 3. 4.
Which actuator has the highest amount of recoverable energy? Should this energy be stored or can it be reused immediately? How difficult, or in other words feasible, is it to recover energy from each of these actuators? How much of the energy sent to the various actuators can be recovered?
It is difficult to give a universally valid answer to all four questions. The amount of recoverable energy depends on the duty cycle and the size of the excavator. To accurately determine these quantities it would be necessary to take a range of differently sized machines and conduct extensive tests. Such a study would be invaluable but has yet to be published. Table 2 summarises measurement data regarding recoverable energy for an 18 t excavator. Four different cycles were analysed. To begin with a simple 90◦ and then 180◦ swing acceleration and braking motion, followed by a 90◦ and 180◦ dig and dump cycle. The data reveals some interesting facts. During a 90◦ swing motion, the machine reaches only seventy per cent of the maximum swing rotational speed as the operator is forced to already start braking around the 60◦ mark in order to come to a stop after 90◦ . During a 180◦ motion, the full swing speed is reached. Due to the quadratic relationship between speed and energy (Equation (9)), a 90◦ rotation has approximately only half as much kinetic energy as a 180◦ rotation. Furthermore, the potential energy x that can be recovered when lowering the boom in a typical dig and dump cycle (D&D) is twice as large as the kinetic energy of the swing at full speed. Despite this, the swing should not be disregarded because, during each D&D cycle, it is accelerated and decelerated twice while the boom is only lowered once. Therefore, in the case of a 180◦ D&D cycle, the swing and boom have the same amount of recoverable energy. For the same cycle, measurements show that ten times less energy can be recovered from the arm actuator. Table 2. Recoverable actuator energy for an 18 t excavator for various duty cycles compared to the recoverable energy x of the boom drive in a typical dig and dump cycle. Swing 90◦
Swing 180◦
Boom D&D
Arm D&D
Swing 90◦ D&D
Swing 180◦ D&D
0.25 x
0.5 x
x
0.1 x
0.5 x
x
For larger machines the amount of boom and swing energy will increase, but it is difficult to judge how their ratio is affected. In the case of the boom, the weight, mg, and the height, ∆h, of the attachment structure increase with the size of the machine. This is not the case for the swing
Energies 2017, 10, 687
15 of 25
drive as the rotational inertia, J, increases with operating weight but the maximum rotational speed, ω, decreases. Figure 14 answers the second question. In a D&D cycle, during swing braking and boom lowering no other actuator requires this amount of power at exactly the same moment. Therefore, the only way to take advantage of the recoverable energy is to store it and then reuse it at a later stage. This takes us to the third question, feasibility. In general, storing energy is not a problem, but the ability to reuse it efficiently and cost effectively determines whether recovery is feasible. If the energy is to be stored hydraulically, two factors are important. First, the design and sizing of the storage device is much Energies 2017, 10, 687 15 of 25 simpler if the actuator pressure during recovery is fairly constant and known beforehand. Secondly, if the recoveryhigher. pressure level is greaterdespite than thethe pressure level during the reuse energy phase, the considerably Consequently, large amount of potential thatstored can oil be can be sent directly to the actuator, making things easier. recovered, reusing it is far more difficult and costly in comparison to the swing drive.
Figure 14. Positive and negative work during dig and dump cycle. Figure 14. Positive and negative work during dig and dump cycle.
Finally, to answer the last question, what portion of the energy delivered to the actuator can be This explains swing is feasible. The to pressure is not only recovered? If this why ratio recovering is small, then it energy does not make sense invest during in suchbraking technologies. An constant but is also higher than the maximum pressure during acceleration. As a result, brake energy integration of the areas under the curves in Figure 14 reveals that approximately half of the energy can easily be storedand andmove usedthe to assist the acceleration phase. The boom far more complex. Due to used to accelerate actuators can in fact be recovered. Put is differently, if all the energy the kinematic arrangement and wide range of lowering speeds, the pressure level during boom down from the actuators could be recovered and reused without any losses, the engine would only have to operation is not constant. Furthermore, the pressure level during boom up operation is usually supply half as much energy. We can conclude that an excavator is very well suited for systems with considerably higher. circuits. Consequently, despite the large amount of potential energy that can be recovered, integrated recovery reusing it is far more and costlyconditions in comparison to described the swing as drive. In summary, thedifficult typical operating can be follows [40]: Finally, to answer the last question, what portion of the energy delivered to the actuator can 1. Average power requirements are then considerably lower thansense peak power requirements. be recovered? If this ratio is small, it does not make to invest in such technologies. 2. Actuator power can be positive (lifting, accelerating) or negative (lowering, The An integration of the areas under the curves in Figure 14 reveals that approximatelydecelerating). half of the energy peak negative power can reach levels similar to the rated engine power. used to accelerate and move the actuators can in fact be recovered. Put differently, if all the energy 3. profiles (pressure and flow rate) of allwithout actuators vary independently of eachonly other. Some fromDemand the actuators could be recovered and reused any losses, the engine would have to actuators require high pressure and low flow rate, while others may require a low pressure and supply half as much energy. We can conclude that an excavator is very well suited for systems with high flow rate. circuits. integrated recovery 4. Idling is common. In summary, the typical operating conditions can be described as follows [40]: Now that the basic concepts regarding how an excavator works and the tasks it performs are 1. Average power requirements are considerably lower than peak power requirements. clear, a couple of definitions and methods to compare losses can be introduced. 2. Actuator power can be positive (lifting, accelerating) or negative (lowering, decelerating). The peak negativeand power can reach levels 4. Defining Efficiency Quantifying Lossessimilar to the rated engine power. Demand profiles (pressure and flow rate) of all actuators vary independently of each other. Some 3. Figure 15 shows a Sankey plot of the energy flow for a 90° dig and dump cycle. The results are actuators require high pressure and low flow rate, while others may require a low pressure and comparable with those published by Sturm for a similar machine [25]. As expected, more than 60% high flow rate. of the energy contained in the fuel is lost through the combustion process. A further 11.7% consumed 4. Idling is common. by the ancillary drives and pump. Approximately half the energy supplied by the pump, EPump , is lost through otherregarding hydraulichow losses, such as pipe friction. Intasks total,itonly aboutare 12% of Now thatthrottling the basic and concepts an excavator works and the performs clear, . Half this energy actually performs useful the total input energy is delivered to the actuators, E a couple of definitions and methods to compare losses Act,can pos be introduced. work on the objects in the surroundings, while the other half EAct,neg is used to raise and accelerate the implement structure. Instead of recovering this portion during lowering and braking actions, it is throttled to tank and is dissipated as heat. In summary, the diagram reveals that approximately a third of the energy supplied by the engine is consumed by the ancillary drives and pump, a third is lost through throttling in the hydraulic system and only the last third actually reaches the actuators.
Energies 2017, 10, 687
16 of 25
4. Defining Efficiency and Quantifying Losses Figure 15 shows a Sankey plot of the energy flow for a 90◦ dig and dump cycle. The results are comparable with those published by Sturm for a similar machine [25]. As expected, more than 60% of the energy contained in the fuel is lost through the combustion process. A further 11.7% consumed by the ancillary drives and pump. Approximately half the energy supplied by the pump, EPump , is lost through throttling and other hydraulic losses, such as pipe friction. In total, only about 12% of the total input energy is delivered to the actuators, EAct, pos . Half this energy actually performs useful work on the objects in the surroundings, while the other half EAct,neg is used to raise and accelerate the implement structure. Instead of recovering this portion during lowering and braking actions, it is throttled to 10, tank Energies 2017, 687and is dissipated as heat. 16 of 25
Figure 15. Sankey through an an 18 18 tt excavator excavator during during 90 90°◦ dig dig and and dump. dump. Figure 15. Sankey plot plot of of energy energy flow flow through
4.1. System Efficiency In summary, the diagram reveals that approximately a third of the energy supplied by the engine Using the the drives hydraulic is defined is consumed bydiagram, the ancillary andsystem pump, aefficiency third is lost throughas: throttling in the hydraulic system and only the last third actually reaches the actuators. EAct, pos ηHyd = (11) EPump 4.1. System Efficiency For the dig and dump cycle a hydraulic efficiency of approximately 49% is given, which is Using the diagram, the hydraulic system efficiency is defined as: actually a very good value for a valve controlled hydraulic system with four actuators and only one supplying pump. The hydraulic system efficiencyEAct, depends on the number of pumps used in the pos (11) ηHyd = system and how the valves are used to distribute the flow among the actuator. Consequently, a EPump typical single circuit Load Sensing system has an efficiency of up to 50%, while dual circuit positive Forsystems the dig and cycle a hydraulic efficiency control can dump reach efficiencies around 60%. of approximately 49% is given, which is actually a very good value for a valve controlled hydraulic system with four actuators and only oneinsupplying Total system efficiency can be defined in two different ways. If we are only interested knowing pump. The hydraulic system efficiency depends on the number of pumps used in the system and how how much of the fuels energy entering the system, EDiesel , is actually delivered to the actuators the are used to distribute the flow among the actuator. Consequently, a typical single circuit EAct,valves pos , then the total gross efficiency is the relevant quantity: Load Sensing system has an efficiency of up to 50%, while dual circuit positive control systems can EAct, pos reach efficiencies around 60%. (12) ηTot, Gross = EDiesel Total system efficiency can be defined in two different ways. If we are only interested in knowing how much of the fuels energy entering the system, EDiesel , is considered, actually delivered to the actuators EAct, pos If the fact that energy can be recovered and reused then the total net efficiency is, relevant: then the total gross efficiency is the relevant quantity: EAct, pos - EAct,neg EAct, net EAct, pos (13) = = (12) Tot, GrossEDiesel EDiesel EDiesel Defining system efficiency in these terms conceals the fact that a machine still consumes fuel when just idling because the individual efficiencies would be calculated to 0% with EAct, pos = 0 disregarding the amount of EDiesel . Instead of discussing efficiencies solely, an evaluation of absolute losses and fuel consumption is more expedient. ηTot, Netη=
Energies 2017, 10, 687
17 of 25
If the fact that energy can be recovered and reused is considered, then the total net efficiency is relevant: EAct, pos − EAct,neg E = Act, net (13) ηTot, Net = EDiesel EDiesel Defining system efficiency in these terms conceals the fact that a machine still consumes fuel when just idling because the individual efficiencies would be calculated to 0% with EAct, pos = 0 disregarding the amount of EDiesel . Instead of discussing efficiencies solely, an evaluation of absolute losses and fuel consumption is more expedient. 4.2. Absolute Losses and Fuel Consumption A far more useful approximation of a machine’s fuel consumption can be derived by combining the Willans approximation (Equation (1)) with the idle power demand relation (Equation (8)) and the definition of hydraulic efficiency (Equation (11)). Because the idle pump losses are already Energies 2017, 10, 687 17 of 25 included in the term PIdle , it is necessary to introduce the differential pump efficiency ∆ηPump , which describes howsetting, much extra engine power must be∆η supplied the pump to produce additional displacement measurements show that cantobe roughly approximated using ahydraulic constant Pump power. Although the pump’s differential efficiency changes slightly depending on the pressure and value of 0.9. displacement setting, measurements show that ∆ηPump can be roughly approximated using a constant PPump, AVE value of 0.9. VDiesel l h = VICE, 0 + 0.22 · PICE = VICE,0 + 0.22 · PIdle + h i ∆ηPump P . . . AVE V Diesel hl = V ICE, 0 + 0.22 · PICE = V ICE, 0 + 0.22· PIdle + Pump, ∆ηPump h i PAct,pos, AVEP . . Act,pos, AVE l (14) = V + 0.22 · m n + (14) = V + 0.22 · m n + VDieselVlDiesel 0 ICE ICE, 0 0 ICE ∆ηPump ηHyd h h h ICE,0 ∆ηPump ηHyd i . . 0.22 V Diesel hl = (V ICE, 0 + 0.22 ·m0 nICE ) + ∆ηPump ηHyd · PAct,pos, AVE 0.22 l VDiesel h = VICE,0 + 0.22 · m0 nICE + ·PAct,pos, AVE ηHyd with To show how useful the equation is, consider an 18∆η t Pump excavator a six-litre diesel engine . operating at 1800 rpm as an example. V ICE, 0 , determined using Equation (2), is approximately 2.9 L/h To show how useful the equation is, consider an 18 t excavator with a six-litre diesel engine and a typical value for m0 is 0.0167 kW/rpm. Assuming an average hydraulic system efficiency of operating at 1800 rpm as an example. VICE,0 , determined using Equation (2), is approximately 2.9 L/h between 30% and 60% results in the following relation and a typical value for m0 is 0.0167 kW/rpm. Assuming an average hydraulic system efficiency of between 30% and 60% results in the following relation l . 9.5 + 0.4074 · PAct,pos, AVE < V Diesel, 1800 rpm < 9.5 + 0.8148 · PAct,pos, AVE (15) 9.5 + 0.4074 · PAct,pos, AVE < VDiesel, 1800 rpm l hh < 9.5 + 0.8148 · PAct,pos, AVE (15) Figure 16 16 shows shows this Figure this inequality inequality graphically. graphically.
Figure 16. Fuel consumption model (engine speed 1800 rpm). Figure 16. Fuel consumption model (engine speed 1800 rpm).
This means that even when just idling and conducting no work, the machine will still consume 9.5 L of fuel per hour. The total amount of fuel consumed due to this phenomenon will be referred to as the fixed fuel consumption VFixed . VFixed =
VDiesel dt
(16)
In the case of a grading cycle with a low average actuator power demand of around 10 kW the fuel consumption only increases by 4.89 L/h compared to idle. This additional amount of fuel
Energies 2017, 10, 687
18 of 25
This means that even when just idling and conducting no work, the machine will still consume 9.5 L of fuel per hour. The total amount of fuel consumed due to this phenomenon will be referred to as the fixed fuel consumption VFixed . VFixed =
Z .
V Diesel dt
(16)
In the case of a grading cycle with a low average actuator power demand of around 10 kW the fuel consumption only increases by 4.89 L/h compared to idle. This additional amount of fuel consumed to actually perform the cycle will be referred to as the variable fuel consumption VVariable . The variable fuel consumption is not affected by the engine operation as the engine’s differential efficiency is constant according to Willans. As shown in Figure 17 only for cycles with average actuator power demands greater Energies 2017, 10, 687 than 20 kW does the variable fuel consumption surpass the fixed fuel consumption. 18 of 25 Energies 2017, 10, 687
18 of 25
Figure 17. Concept of fixed and variable fuel consumption. Figure 17. 17. Concept Concept of of fixed fixed and and variable variable fuel fuel consumption. consumption. Figure
4.3. Model Validation 4.3. Model Model Validation Validation 4.3. The very simplified model behind Equations (14) and (15) does not consider dynamic effects as The very very simplified simplifiedmodel modelbehind behind Equations(14) (14)and and(15) (15) does not consider dynamic effects The does not consider dynamic effects as frequent transient loading conditions,Equations which further increase fuel consumption. Despite this, as frequent transient loading conditions,which whichfurther furtherincrease increase fuel fuel consumption. consumption. Despite Despite this, this, frequent transient loading conditions, measurements conducted on a real machine correlate well with the predicted theoretical values. measurements conducted conducted on on aa real real machine machine correlate correlate well well with with the the predicted predicted theoretical theoretical values. values. measurements Figure 18 proves the good comparability of the measured fuel rate and the introduced fuel Figure 18 proves the good comparability of the measured fuel rate and the introduced fuel consumption Figure 18 proves good comparability of the measured fuel rate and was the calculated introduced consumption modelthe (Equation (14)) with a hydraulic efficiency of 49%, which forfuel the model (Equation (14))(Equation with a hydraulic efficiency of 49%, which of was calculated for the dig andfor dump consumption model (14)) with a hydraulic efficiency 49%, which was calculated the dig and dump cycle in Section 4.2. cycle in Section 4.2. in Section 4.2. dig and dump cycle
Figure 18. Measurements of load sensing (LS) system compared to fuel consumption model. Figure 18. Measurements of load sensing (LS) system compared to fuel consumption model. Figure 18. Measurements of load sensing (LS) system compared to fuel consumption model.
Fuel rate fluctuations around the idealized Willans line are caused by a varying hydraulic Fuel rate fluctuations around the idealized Willans line are caused by whereby a varying hydraulic efficiency during the cycle and the mentioned transient loading of the engine, more fuel is efficiency during the cycle and the mentioned transient loading of the engine, whereby more fuel is consumed to prevent the engine from stalling. Averaged values of fuel rate and positive actuator consumed to prevent the engine Averaged values of fuel rate and power for the dig and dump cyclefrom were stalling. calculated and are also plotted in Figure 18.positive The lowactuator average power for the dig and dump cycle were calculated and are also plotted in Figure 18. The low average power demand compared to the maximum power demand becomes apparent. This analysis was power demand compared to the maximum power demand becomes apparent. This analysis was
Energies 2017, 10, 687
19 of 25
Fuel rate fluctuations around the idealized Willans line are caused by a varying hydraulic efficiency during the cycle and the mentioned transient loading of the engine, whereby more fuel is consumed to prevent the engine from stalling. Averaged values of fuel rate and positive actuator power for the dig and dump cycle were calculated and are also plotted in Figure 18. The low average power demand compared to the maximum power demand becomes apparent. This analysis was done for a variety different load cycles (see Figure 19). Energies 2017, 10,of687 19 of 25
Figure 19. Validation of fuel consumption model. Figure 19. Validation of fuel consumption model.
Most of the cycles reach hydraulic efficiencies in the range of 30–50%. In contrast to that the air Most of the cycles reach hydraulic in the range of In contrast to that the grading cycle, which simulates a commonefficiencies grading cycle (cf. Figure 10)30–50%. but without any external loads, air grading cycle, which simulates a common cycleefficiency (cf. Figure butThis without reveals a hydraulic efficiency of 22% and a grading total gross of10) 5%. resultany is external a direct loads, reveals a hydraulic efficiency of 22% and a total gross efficiency of 5%. This result is aduring direct consequence of the single circuit hydraulic system used in wheeled excavators. At one stage consequence of the single circuit hydraulic system used in wheeled excavators. At one stage during this cycle the boom is raised and simultaneously the arm cylinder is extended rapidly. Hence the this cycle the boom is by raised simultaneously the arm cylinder is extended rapidly. Hence the system pressure is set the and boom’s piston pressure, which is considerably higher (approximately system pressure is set by the boom’s piston pressure, which is considerably higher (approximately 110 bar) than the required pressure to supply the arm cylinder. In combination with a high flow 110 bar) than required pressure 200 to supply arm cylinder. combination with high flow demand of thethe arm (approximately L/min) the throttling losses ofInabout 35 kW can be aobserved at demand of the arm (approximately 200 L/min) throttling losses of about 35 kW can be observed at the the inlet edge of the arm control valve. In the case of a real grading cycle the load and accordingly inlet edge of the armiscontrol In thethe case of drive, a real grading the load and throttling accordingly the the pressure, which neededvalve. to supply arm increasescycle leading to lower losses pressure, which is needed to supply the armofdrive, to lower and and finally to a higher hydraulic efficiency 30%.increases Based onleading these results, anthrottling averagedlosses hydraulic finally to a higher hydraulic efficiency of 30%. Based on these results, an averaged hydraulic efficiency efficiency of 40% is assumed hereafter, see Equation 17. of 40% is assumed hereafter, see Equation 17. VDiesel, 1800 rpm, η = 0,4 l h = 9.5 + 0.6111 · PAct,pos, AVE (17) Hyd . l V Diesel, 1800 rpm,ηHyd = 0,4 = 9.5 + 0.6111 · PAct,pos, AVE (17) h 5. Predicting Fuel Consumption Improvements 5. Predicting Fuel Consumption Improvements Now that the loss mechanisms occurring in an excavator are clear, we can take the next step and Nowways that the loss mechanisms occurring into anpredict excavator clear, we can take the next and consider to reduce fuel consumption and the are level of improvement with thestep help of consider ways to reduce fuel consumption and to predict the level of improvement with the help of this simple model using Willans lines. this simple model using Willans lines. 5.1. Saving Fuel by Lowering Hydraulic Losses 5.1. Saving Fuel by Lowering Hydraulic Losses The majority of research in the field of energy efficient excavators has dealt with the The majority of research in the hydraulic field of energy efficient excavators has dealt with development development of methods to lower losses. A number of publications have the investigated the of methods to lower hydraulic losses. A number of publications have investigated the ability of ability of different oils to improve hydraulic efficiency [1,41,42]. However, what most of these studies different oils to improve hydraulic efficiency [1,41,42]. However, what most of these studies have have failed to discuss is the direct impact of these changes to the hydraulic systems on fuel failed to discuss is theEquation direct impact of these changes to thereduction hydraulicofsystems on fuel consumption. consumption. Using (18), the fuel consumption a machine operating with a Using Equation (18), the fuel consumption reduction of a machine operating with a theoretically theoretically lossless hydraulic system can be studied: lossless hydraulic system can be studied: VDiesel, 1800rpm, η = 1 l h = 9.5 + 0.2444 · PAct,pos, AVE (18) Hyd Figure 20 reveals that even if it were possible to design a system capable of completely eliminating throttling, fuel consumption could be reduced by no more than 30% for a dig and dump cycle.
Energies 2017, 10, 687
20 of 25
.
V Diesel, 1800 rpm,ηHyd = 1
l = 9.5 + 0.2444 · PAct,pos, AVE h
(18)
Figure 20 reveals that even if it were possible to design a system capable of completely eliminating 20 of 25 throttling, fuel consumption could be reduced by no more than 30% for a dig and dump cycle. 20 of 25 Energies 2017, 10, 687 Energies 2017, 10, 687
Figure 20. Maximum theoretical fuel consumption reduction with a lossless hydraulic system. Figure 20. Maximum theoretical fuel consumption reduction with a lossless hydraulic system. Figure 20. Maximum theoretical fuel consumption reduction with a lossless hydraulic system.
For other cycles with a lower average positive actuator power this theoretical value is even For other cycles with a lower average positive actuator power this theoretical value is even lower. Forhydraulic other cycles with capable a lower average positive actuatorlosses powerto this is even lower. lower. A circuit of reducing throttling a theoretical minimum value is a displacement A hydraulic circuit capable of reducing throttling losses to a minimum is a displacement A hydraulic reducing throttling losses to aa single minimum is a displacement controlled system. circuit Insteadcapable of usingof valves to control actuator motion, hydraulic unitcontrolled supplies system. Instead of using valves to control actuator motion, a single hydraulic unit supplies each controlled system. Instead of using valves to control actuator motion, a single hydraulic unit supplies each actuator with the required power individually. The actuators velocity is controlled by regulating actuator with the required power individually. The actuators velocity is controlled by regulating the each actuatordisplacement. with the required power individually. The actuators velocity is controlled regulating the pump’s Unfortunately, the considerable higher costs as well as thebylow system pump’s displacement. Unfortunately, the considerable higher costs as well as the low system damping the pump’s displacement. Unfortunately, the prevented considerable higher as the and low easy system damping associated with these systems have them fromcosts beingasanwell economic to associated with these systems have prevented them from being an economic and easy to control damping associatedto with these systems have prevented them frompromising, being an economic to control alternative valve controlled systems. Another, more approachand to easy reduce alternative to valve systems. Another, more promising, approach to reduce losses control alternative to valve controlled systems. Another, more promising, reduce throttling losses is controlled independent metering valves, which decouple the meter inapproach andthrottling the to meter out is independent metering valves, decouple thewhich meterto inmaintain and thethe meter outin edges. As even mentioned throttling is independent metering valves, decouple meter andand the metermore out edges. As losses mentioned before the which outlet edge is necessary controllability before the outlet edge is necessary to maintain controllability and even more important to prevent edges. As mentioned the outlet is necessary to controllable maintain controllability and even morea important to preventbefore runaway loads. edge An independently meter out edge enables runaway loads. An independently controllable meter out on edge enables reduction of edge throttling losses important tothrottling prevent runaway An independently controllable meter out a reduction of losses to aloads. minimum depending the actual aload pressure of theenables actuator. to a minimum depending on the actual load pressure of the actuator. With such a system hydraulic reduction of throttling losses to a minimum depending on the actual load pressure of the actuator. With such a system hydraulic efficiencies of about 70% can be achieved. However, as shown in efficiencies about 70% can beefficiencies achieved. However, asdig shown Figure 21, thethereby reduction in fuel With such aofsystem hydraulic of about 70% can be in achieved. However, as limited shown in Figure 21, the reduction in fuel consumption for the and dump cycle is to consumption for the dig and dump cycle is thereby limited to approximately 17%. Figure 21, the 17%. reduction in fuel consumption for the dig and dump cycle is thereby limited to approximately approximately 17%.
Figure 21. Theoretical fuel consumption reduction of a hydraulic system with 70% efficiency. Figure 21. Theoretical fuel consumption reduction of a hydraulic system with 70% efficiency. Figure 21. Theoretical fuel consumption reduction of a hydraulic system with 70% efficiency.
Solutions solely focused on the hydraulic subsystem, without any consideration of the fixed fuel Solutions are solely focusedvery on the hydraulic subsystem, without any consideration of the fixed fuel consumption, therefore limited. consumption, are therefore very limited. 5.2. Saving Fuel by Lowering Idle Losses 5.2. Saving Fuel by Lowering Idle Losses Although it is not often mentioned explicitly, the concept of idle consumption is actually behind it is not often mentioned explicitly, the concept of idle consumptionengine is actually behind manyAlthough of the fuel saving methods discussed in literature. Engine downspeeding, downsizing many of the fuel savingall methods discussed literature. downspeeding, downsizing and start–stop systems improve machine in efficiency byEngine lowering idling losses. engine Downspeeding not
Energies 2017, 10, 687
21 of 25
Energies 2017, 10, 687
21 of 25 Solutions solely focused on the hydraulic subsystem, without any consideration of the fixed fuel consumption, are therefore very limited. (Equation (2)), but also lowers the parasitic power demanded by the ancillary drives (Equation (8)). For using the sameIdle simplified 5.2. example, Saving Fuel by Lowering Losses model from above, by reducing engine speed from 1800 rpm down to 1200 rpm, the fuel consumption relation becomes: Although it is not often mentioned explicitly, the concept of idle consumption is actually behind l literature. VDiesel, · PAct,pos, (19) many of the fuel saving methods discussed Engine downspeeding, engine downsizing 1200rpm in AVE h = 6.2 + 0.489 and start–stop systems all improve machine efficiency by lowering idling losses. Downspeeding Figure 22 shows the relation graphically. Especially for cycles with low average power demands, not only lowers the friction losses in the engine itself, which are a quadratic function of the speed this measure is very effective, more so than any optimisation of the hydraulic system. The efficiency (Equation (2)), but also lowers the parasitic power demanded by the ancillary drives (Equation (8)). improvements resulting from downspeeding have been shown in numerous studies [20,43,44]. What For example, using the same simplified model from above, by reducing engine speed from 1800 rpm these studies have also shown is that a reduced engine speed also lowers productivity because the down to 1200 rpm, the fuel consumption relation becomes: engine can no longer supply the same amount of power as at the higher speed. Downsizing, i.e., using. a smaller engine l with less power, is also discussed frequently, but is very V Diesel, 1200rpm = 6.2 + 0.489 · PAct,pos, AVE (19) difficult to implement in an excavator. For hduty cycles like travelling and drilling with an attachment, in which full power is demanded for prolonged periods of time a downsized engine will just not have Figure 22 showsAlthough the relation graphically. for than cycles with low average power demands, the required power. these cycles areEspecially less frequent others, the machine must be capable this measure is very effective, more so than any optimisation of the hydraulic system. The efficiency of performing them. Start–stop systems are particularly interesting, as the engine is completely improvements resulting from downspeeding have This beenisshown in numerous [20,43,44]. switched off meaning absolutely no fuel is consumed. especially efficient forstudies cars in city traffic What these studies have also shown is that a reduced engine speed also lowers productivity because that are frequently stopping at traffic lights. Implementing this technology in construction machinery theconsiderably engine can no longer supply the same amount of power as at the higher speed. is more difficult.
Figure 22. Theoretical fuel consumption reduction possible with engine downspeeding. Figure 22. Theoretical fuel consumption reduction possible with engine downspeeding.
Other methods to lower idle power demand, which are not directly related to the engine, include Downsizing, using a smaller engine withand lessusing power, is also discussed frequently, but is very removing ancillaryi.e., drives from the engine shaft decentralised electric drives, for example difficult to implement in an excavator. For duty cycles like travelling and drilling with an attachment, for the cooling fan pump. in which full power is demanded for prolonged periods of time a downsized engine will just not haveSaving the required power. Although these cycles are less frequent than others, the machine must be 5.3. Fuel through Energy Recovery capable of performing them. Start–stop systems are particularly interesting, as the engine is completely Various methods of recovering and reusing both boom and swing energy have been proposed switched off meaning absolutely no fuel is consumed. This is especially efficient for cars in city traffic in literature and can now be found in various series machines available on the market [45]. An that are frequently stopping at traffic lights. Implementing this technology in construction machinery important issue to address is the potential of energy recovery methods to lower fuel consumption. is considerably more difficult. To do so the fuel consumption model is extended to include negative actuator power. If a recovery Other methods to lower idle power demand, which are not directly related to the engine, include system with an efficiency of ηRec were installed the engine would be required to supply less power removing ancillary drives from the engine shaft and using decentralised electric drives, for example and thecooling averagefan fuel consumption would be for the pump. VDiesel, 1800 rpm l h = 9.5 + 0.6111 · (PAct,pos, AVE - ηRec · PAct,neg, AVE ) (20) Measurements show that the average negative actuator power is always less than or equal to 50% of the average positive actuator power. For cycles such as dig and dump the maximum of 50% is reached but in the case of trenching this value goes down to approximately 30%. This can be expressed with the following inequality:
Energies 2017, 10, 687
22 of 25
5.3. Saving Fuel through Energy Recovery Various methods of recovering and reusing both boom and swing energy have been proposed in literature and can now be found in various series machines available on the market [45]. An important issue to address is the potential of energy recovery methods to lower fuel consumption. To do so the fuel consumption model is extended to include negative actuator power. If a recovery system with an efficiency of ηRec were installed the engine would be required to supply less power and the average fuel consumption would be .
V Diesel, 1800 rpm
l = 9.5 + 0.6111 · ( PAct,pos, AVE − ηRec · PAct,neg, AVE ) h
(20)
Measurements show that the average negative actuator power is always less than or equal to 50% of the average positive actuator power. For cycles such as dig and dump the maximum of 50% is reached but in the case of trenching this value goes down to approximately 30%. This can be expressed with the following inequality: Energies 2017, 10, 687
PAct,neg, AVE ≤ 0.5 · PAct,pos, AVE
22 of 25
(21)
Inserting Equation (20) into Equation a AVE recovery efficiency of 100%, gives(21) the PAct,neg, (21), 0.5 · PAct,pos, AVE ≤assuming minimum theoretical fuel consumption of a perfect recovery system, capable of recovering boom, Inserting Equation (20) into Equation (21), assuming a recovery efficiency of 100%, gives the swing and arm energy (see Figure 23). minimum theoretical fuel consumption of a perfect recovery system, capable of recovering boom, . swing and arm energy (see l Figure 23). = 9.5 + 0.6111 · 0.5 · PAct,pos,AVE = 9.5 + 0.3056 · PAct,pos,AVE (22) V Diesel, 1800 rpm,ηrec = 1 h VDiesel, 1800 rpm, η = 1 l h = 9.5 + 0.6111 · 0.5 · PAct,pos,AVE = 9.5 + 0.3056 · PAct,pos,AVE (22) rec
Figure 23. Maximum theoretical fuel consumption reduction possible with energy recovery. Figure 23. Maximum theoretical fuel consumption reduction possible with energy recovery.
5.4. Saving Fuel through Holistic Approach 5.4. Saving Fuel through Holistic Approach The only way to drastically lower fuel consumption is to follow a holistic approach [40,46]. This The only way to drastically lower fuel consumption is to follow a holistic approach [40,46]. means improving the hydraulics, enabling energy recovery and lowering idle fuel consumption. For This means improving the hydraulics, enabling energy recovery and lowering idle fuel consumption. a machine operating at 1200 rpm, with a lossless hydraulic system and optimal energy recovery, the For a machine operating at 1200 rpm, with a lossless hydraulic system and optimal energy recovery, minimum theoretical fuel consumption can be calculated as follows: the minimum theoretical fuel consumption can be calculated as follows: VDiesel, 1200 rpm, rec, min l h = 6.2 + 0.2444 · 0.5 · PAct,pos,AVE = 6.2 + 0.1222 · PAct,pos,AVE (23) . l V Diesel, 1200 rpm, rec, min = 6.2 + 0.2444 · 0.5 · PAct,pos,AVE = 6.2 + 0.1222 · PAct,pos,AVE (23) h this results in both a parallel displacement and a change in gradient of As shown in Figure 24, the consumption line. Theoretically, consumption can be reduced by up 59% for the dig and dump As shown in Figure 24, this results in both a parallel displacement and a change in gradient of the cycle. In practice, this value cannot be reached but it is helpful to have a theoretical limit. consumption line. Theoretically, consumption can be reduced by up 59% for the dig and dump cycle. In practice, this value cannot be reached but it is helpful to have a theoretical limit.
minimum theoretical fuel consumption can be calculated as follows: VDiesel, 1200 rpm, rec, min l h = 6.2 + 0.2444 · 0.5 · PAct,pos,AVE = 6.2 + 0.1222 · PAct,pos,AVE
(23)
As shown in Figure 24, this results in both a parallel displacement and a change in gradient of the consumption Energies 2017, 10, 687 line. Theoretically, consumption can be reduced by up 59% for the dig and dump 23 of 25 cycle. In practice, this value cannot be reached but it is helpful to have a theoretical limit.
Figure 24. Maximum theoretical fuel consumption reduction using engine downspeeding, lossless Figure 24. Maximum theoretical fuel consumption reduction using engine downspeeding, lossless hydraulics and energy recovery. hydraulics and energy recovery.
6. Conclusions 6. Conclusions A detailed examination of measurement data shows that the fuel consumption of an excavator A detailed examination of measurement data shows that the fuel consumption of an excavator can can be divided into a fixed and variable component. The fixed component quantifies the amount of be divided into a fixed and variable component. The fixed component quantifies the amount of fuel fuel the machine consumes if it was just left at idle and did not move for the entire cycle time. The the machine consumes if it was just left at idle and did not move for the entire cycle time. The variable variable component specifies the additional amount of fuel consumed in order to actually perform component specifies the additional amount of fuel consumed in order to actually perform the task. Depending on the cycle the ratio of fixed to variable fuel consumption changes. In cycles with low average power demands, for example grading, the fixed consumption is actually greater than the variable consumption. Consequently, lowering the total fuel consumption of a machine, requires a consideration of both the fixed and variable terms. The model introduced in this paper helps assess the potential of certain measures regarding their ability to lower fuel consumption a priori. An interesting result is that a hydraulic system with 100% efficiency can only reduce the fuel consumption during grading by approximately 25%. This is because improvements in the hydraulic system efficiency only lower the variable consumption, not the fixed consumption. Similarly, a machine capable of recovering all available boom potential and swing kinetic energy can theoretically only reduce fuel consumption by a maximum of around 30%. Consequently, thinking only in terms of hydraulic efficiency is rather misleading. A more holistic approach, considering engine operation and other important factors, is far better suited. Author Contributions: Milos Vukovic and Roland Leifeld performed the measurements and analysed the data. All authors contributed to the writing of the paper. Conflicts of Interest: The authors declare no conflict of interest.
References 1. 2. 3. 4.
5.
Abekawa, T.; Tanikawa, Y.; Hirosawa, A. Introduction of Komatsu Genuine Hydraulic Oil KOMHYDRO HE; Komatsu Technical Report; Yumpu: Komatsu, Japan, 2010; Volume 56, No. 163. Ohira, S.; Suehiro, M.; Ota, K.; Kawamura, K. Use of emission rights for construction machinery to help prevent global warming. Hitachi Rev. 2013, 62, 123–130. Bloomfield, L.A. How Things Work: The Physics of Everyday Life; John Wiley & Sons: Inc.: Hoboken, NJ, USA, 2013. Pragmatic Efficiency Limits for Internal Combustion Engines, Readout Guest Forum. 2014. Available online: http://www.horiba.com/publications/readout/article/pragmatic-efficiency-limits-for-internalcombustion-engines-31871/ (accessed on 7 October 2016). Hanlon, M. Most Powerful Diesel Engine in the World, Gizmag. Available online: http://www.gizmag. com/go/3263/ (accessed on 7 October 2016).
Energies 2017, 10, 687
6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23.
24. 25. 26.
27. 28. 29. 30. 31. 32.
24 of 25
Frei, B. Regelung Eines Elektromechanischen Getriebes für Hybridfahrzeuge. Ph.D. Thesis, Technical University Chemnitz, Chemnitz, Germany, 2006. Yang, J.; Quan, L.; Yang, Y. Excavator energy-saving efficiency based on diesel engine cylinder deactivation technology. Chin. J. Mech. Eng. 2012, 25. [CrossRef] Liebl, J.; Lederer, M.; Rohde-Brandenburger, K.; Biermann, J.-W.; Roth, M.; Schäfer, H. Energiemanagement im Kraftfahrzeug; Springer: Wiesbaden, Germany, 2014. Rohde-Brandenburger, K. Was bringen 100 kg Gewichtsreduzierung im Verbrauch? In ATZ; Springer: Wiesbaden, Germany, 2013; Volume 115. Guzzella, L.; Sciaretta, A. Vehicle Propulsion Systems; Springer: Berlin/Heidelberg, Germany, 2013. Wilde, A. Eine Modulare Funktionsarchitektur für Adaptives Und Vorausschauendes Energiemanagement in Hybridfahrzeugen. Ph.D. Thesis, Technical University Munich, München, Germany, 2008. Smil, V. Energy: A Beginner’s Guide; Oneworld Publications: London, UK, 2006. Chiatti, G.; Chiavola, O.; Falcucci, G. Soot Morphology Effects on DPF Performance; SAE Technical Paper 2009-01-1279; SAE: Warrendale, PA, USA, 2009. [CrossRef] Falcucci, G. A lumped parameter model for diesel soot morphology evaluation and emission control. Proc. Instit. Mech. Eng. Part D J. Automob. Eng. 2012, 226, 987–998. [CrossRef] Filla, R. Getting Real about Test Cycles; SAE International Off-Highway Engineering: Warrendale, PA, USA, 2013; ISSN: 1528-9702. Terdich, N. Impact of Electrically Assisted Turbocharging on the Transient Response of an Off-Highway Diesel Engine. Ph.D. Thesis, Imperial College London, London, UK, 2015. Hanson, P.A.; Lindgren, M.; Nordin, M.; Pettersson, O. A methodology for measuring the effects of transient loads on the fuel efficiency of agricultural tractors. Appl. Eng. Agric. 2003, 19. [CrossRef] Lindgren, M.; Hansson, P.-A. Effects of transient conditions on exhaust emissions from two non-road diesel engines. Biosyst. Eng. 2004, 87, 57–66. [CrossRef] Forschungsdialog: Erfolgsprojekt STEAM, O+P Ölhydraulik und Pneumatik 07/08; Vereinigte Fachverlage: Mainz, Germany, 2016. Kunze, G.; Göhring, H.; Jacob, K. Baumaschinen: Erdbau-Und Tagesbaumaschinen; Vieweg Verlag: Wiesbaden, Germany, 2002. Holländer, C. Untersuchungen zur Beurteilung und Optimierung von Baggerhydrauliksystemen. Ph.D. Thesis, TU Braunschweig, Braunschweig, Germany, 1988. Herfs, W. LUDV-Steuerungen für Bagger. Europäische Mobiltagung, Lohr a. Main, Mannesmann Rexroth AG; Presentation (RD 00 207/10.97); Rexroth: Lohr, Germany, 1997. N.N. Caterpillar M318D Wheeled Excavator Brochure. Available online: http://www.cat.com/ en_US/products/new/equipment/wheel-excavators/wheel-excavators/18115489.html (accessed on 7 October 2016). Kawasaki K3VL Axial piston pumps for open circuits in mobile. In Industrial and Marine Applications; Datasheet: Kawasaki, Japan, 2014. Sturm, C. Bewertung der Energieeffizienz von Antriebssystemen mobiler Arbeitsmaschinen am Beispiel Bagger. Ph.D. Thesis, Karlsruhe Institute of Technology, Karlsruhe, Germany, 2015. Vukovic, M.; Sgro, S.; Murrenhoff, H. STEAM—Hydraulic design for engine integration. O + P : Fluidtechnik für den Maschinen- und Anlagenbau 2015, 59. Available online: https://publications.rwth-aachen.de/record/ 466125 (accessed on 29 March 2017). Achten, P. Dynamic high-frequency behaviour of the swash plate in a variable displacement axial piston pump. J. Syst. Control Eng. 2013, 227, 1–12. [CrossRef] Lux, J.; Murrenhoff, H. Experimental loss analysis of displacement controlled pumps. In Proceedings of the 10th International Fluid Power Conference, Dresden, Germany, 8–10 March 2016. Zhang, J. Pressure compensated metering valve for EH systems. In Proceedings of the National Conference on Fluid Power, Las Vegas, NV, USA, 12–14 March 2008. Murrenhoff, H. Servohydraulik—Geregelte Hydraulische Antriebe, 4th ed.; Shaker: Aachen, Germany, 2012. LiDAT the Complete Comprehensive Data Collection and Tracking System from Liebherr; Liebherr Product Brochure; Liebherr: Kirchdorf, Germany, 2016. Bolz, G. Konzepte zur Erhöhung der Energieeffizienz von Mobilbaggern. In Proceedings of the 5th Fachtagung Baumaschinentechnik, Dresden, Germany, 20–21 September 2012.
Energies 2017, 10, 687
33.
34. 35. 36. 37. 38.
39. 40. 41.
42.
43. 44. 45. 46.
25 of 25
Fecke, M. Standardisierung definierter Lastzyklen und Messmethoden zur Energieverbrauchsermittlung von Baumaschinen. In Proceedings of the 5th Fachtgaung Baumaschinentechnik, Dresden, Germany, 17–19 September 2015. Steindorff, K. Energierückgewinnung am Beispiel eines ventilgesteuerten hydraulischen Antriebs. Ph.D. Thesis, TU Braunschweig, Braunschweig, Germany, 2010. Sgro, S.; Murrenhoff, H. Energierückgewinnungssysteme für Baggerausleger—Eine Systematische Übersicht der Vorhandenen Lösungsmöglichkeiten; O+P 10/2010; Vereinigte Fachverlage: Mainz, Germany, 2010. Amrhein, J.; Neumann, U. PRB—Regeneration of potential energy while boom-down. In Proceedings of the 8th International Conference on Fluid Power, Dresden, Germany, 26–28 March 2012. Thompson, B.; Yoon, H.; Kim, J.; Lee, J. Swing Energy Recuperation Scheme for Hydraulic Excavators; SAE Technical Paper 2014-01-2402; SAE: Warrendale, PA, USA, 2014. [CrossRef] Hießl, A.; Scheidl, R. Energy consumption and efficiency measurements of different excavators—Does hybridization pay? In Proceedings of the ASME/BATH Symposium on Fluid Power and Motion Control, Chicago, IL, USA, 12–14 October 2015. Joo, C.; Stangl, M. Application of power regenerative boom system to excavator. In Proceedings of the 10th International Conference on Fluid Power, Dresden, Germany, 8–10 March 2016. Vukovic, M.; Sgro, S.; Murrenhoff, H. STEAM—A holistic approach to designing excavator systems. In Proceedings of the 9th International Fluid Power Conference, Aachen, Germany, 24–26 March 2014. Korane, K. How Hydraulic Fluids Affect Energy Efficiency, Mobile Hydraulic Tips. Available online: http://www.mobilehydraulictips.com/how-hydraulic-fluids-affect-energy-efficiency/ (accessed on 7 October 2016). Michael, P.W.; Mettakadapa, S. Bulk Modulus and Traction Effects in an Axial Piston Pump and a Radial Piston Motor. In Proceedings of the 10th International Fluid Power Conference, Dresden, Germany, 8–10 March 2016. Forche, J. Antriebsstrangmanagement eines Hydraulikbaggers. Ph.D. Thesis, Technical University Braunschweig, Braunschweig, Germany, 2007. Ng, F.; Harding, J.A.; Glass, J. An eco-approach to optimise efficiency and productivity of a hydraulic excavator. J. Clean. Prod. 2016, 112, 3966–3976. [CrossRef] Inderelst, M. Efficiency Improvements in Mobile Hydraulic Systems, Ph.D. Thesis, RWTH Aachen University, Aachen, Germany, 2013. Leifeld, R.; Vukovic, M.; Murrenhoff, H. Hydraulic hybrid architecture for excavators. ATZoffhighway Worldw. 2016, 9, 44–49. [CrossRef] © 2017 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
