ls

lanthile

the

forrall

tonts

wertionweronsuire

a, ahastedon-

thethet inme

itial

CIRP Annals - Manufacturing Technology 61 (2012) 43–46

and

sing

ng a

logy

died

hine

IRP.

Development of an energy consumption monitoring procedure for machine too

Thomas Behrendt a, Andre Zein a,b, Sangkee Min (2)c,*a TU Braunschweig, Braunschweig, Germanyb Joint German-Australian Research Group ‘Sustainable Manufacturing and Life Cycle Management’, Sydney, Australiac Laboratory for Manufacturing and Sustainability (LMAS), University of California, Berkeley, CA, USA

1. Introduction

Energy reduction strategies are increasingly important with theconstant increase in electricity costs and the rising environmentalawareness of both manufacturers and customers. Machine toolshighly contribute to energy consumption in the industrial sector,which is the most energy-consuming sector in the U.S. with a shareof 31% in 2010. At the same time, costs for energy have increased byalmost 70% since the late 1990s [1]. The high environmentalimpact of these products was also declared by the EuropeanCommission within its ‘‘Energy using products’’ directive, whichwas created for improving the environmental performance ofproducts with excessive energy consumption [2].

Most research provided only little insight into how the datamining was done. This results in challenging interpretations whencomparing data of different machines. Thus, the development ofeffective optimization strategies that are valid for the majority ofmachine tools urgently requires standardized test procedures. Thework presented here addresses this problem by deriving for-malized assessment procedures and developing a novel standardtest piece.

2. Energy consumption of machine tools

Most studies provided in literature focus on measuring thepower consumption of different types of machine tools as a basisfor identifying optimization potentials. A commonly acceptedenergy breakdown by Dahmus and Gutowski divides the power

have a constant power demand (e.g. spindle motor and coopumps). The production mode represents the power demand wremoving material. This portion is variable and dependent onload applied to the machine [3].

Numerous studies showed that the power necessary

removing the actual material has only little impact on the oveenergy consumption [4,5]. Thus, different approaches aimedreduce the constant part by either improving specific componeor by reducing the overall cycle time [6,7]. Influences on podemand were identified, such as process parameters [8], selecof tooling [9], and workpiece material [10]. Hence, pomeasurements are highly dependent on a variety of preconditithat directly affect the obtained data and therefore reqstandardized test procedures.

3. Standards in energy evaluation of machine tools

Based on the above-mentioned variability of power datcontroversial discussion about standardized procedures

recently evolved. Dornfeld noted that unambiguous data creathrough standards is necessary to accomplish reliable envirmental evaluations [11]. In addition, Kuhrke et al. emphasizedneed of further standardization efforts and point out that

definition of such procedures is essential for each componenorder to ensure the calculation of energy consumption on the sabasis [12].

The Japanese Standards Association (JSA) published an in

A R T I C L E I N F O

Keywords:

Energy

Monitoring

Machine tool

A B S T R A C T

A systematic method to assess energy consumption of machine tools for comparable analysis of data

to accurately evaluate the energy efficiency of various machine tools is necessary with increa

interests in green manufacturing. This paper proposes a novel and coherent methodology by presenti

detailed description of different test procedures based on standardized workpieces. The methodo

was successfully applied to nine machining centers. Energy consumption characteristics of the stu

machine tools are compared and the potential of using the obtained data for energy labeling of mac

tools is discussed.

� 2012 C

Contents lists available at SciVerse ScienceDirect

CIRP Annals - Manufacturing Technology

journal homepage: http: / /ees.elsevier.com/cirp/default .asp

hisre-

g at al.

by

demand of machine tools into three different modes: idle mode,run-time mode, and production mode. In idle mode, the machine isready for production and components such as the operation paneland fans accumulate to a constant power demand. During run-timemode, further auxiliaries are activated, which, once turned on,

ntsrers

* Corresponding author.

0007-8506/$ – see front matter � 2012 CIRP.

http://dx.doi.org/10.1016/j.cirp.2012.03.103

approach towards standardized power data in 2010 [13]. Tmethodology describes different procedures for power measuments of the three previously described modes, includinstandard workpiece for three-axis milling machines. Abele epointed out that this workpiece, however, could only be usedcompanies for comparing their energy efficiency improvemewithout being able to compare among different manufactu[14].

Adeve(ISOtoolthe

desienerof mimpdefi

4. D

Tmetsequ

Imacidenthatworand

subs(0.1accofor

resp

4.1.

Fdemmacshalwhi

4.2.

Tpowprovactivfor

conssuggmovd–a–

Finclu8, anwasratedoufrom

TableStand

#

1

2

3

4

5

Sta

6

7

8

9a O

T. Behrendt et al. / CIRP Annals - Manufacturing Technology 61 (2012) 43–4644

nother work on standardized energy data is currentlyloped by the International Organization for Standardization) within the framework ‘‘Environmental evaluation of machines’’ (ISO/NP 14955) [15]. The scope of this standard is to defineenvironmental performance of machine tools regarding theirgn with a focus on metal cutting and forming machines. Thegy consumption, associated CO2-emissions, and consumptionaterials are taken into account as the main environmental

acts [16]. However, methods and procedures are yet to bened.

evelopment of a standardized test procedure

he developed test procedure is composed of three assessmenthods revising the energy demand in idle mode, in operationalences, and machining operations.

n order to ensure a scalability of the test procedures, 232hine tools of four major manufacturers were surveyed thattified the work area (i.e. X- times Y-travel) as the parameter

correlates most accurately with the size of a machine tool. Thek area (A) of the studied machine tools varies between 0.02 m2

10 m2. Three main groups of machine tool sizes areequently classified: small (A � 0.1 m2), medium

m2 < A � 1 m2), and large (A > 1 m2). Small machine toolsunt for only 2%, whereas medium-sized machine tools account58% and large machine tools for 40% of 232 machines,ectively.

Standby power

irst, a ‘‘standardized start-up procedure’’ evaluates the powerand during idle mode, which is independent on the size of thehine and actual material processing. Power measurementsl be commenced according to the sequence given in Table 1,ch ensures an accurate measurement of standby power.

Component power

he ‘‘component cycle’’ was derived to assess in detail theer consumed by the remaining main components, whichided a process-induced power demand. Each peripheral wasated for three times, while the machine dwelled in each mode10 s to achieve statistically significant data. The powerumption of the spindle at different speeds was measured asested by the JSA (Fig. 1(left)). The drives were actuated bying the spindle along a cubic motion pattern following a–b–c–e–h–d–f–h–a–g–b–a (Fig. 1(right)).

ive different feed rates were used during the motion pattern,ding maximum feed rate (fmax), rapid traverse rate and 1/4, 1/d 1/2 fmax. If featured by the machine, the worktable or spindle

rotated around the 4th and 5th axis (including rapid traverse). The travel distance between the vertices of the cube wasbled from small (100 mm) to medium (200 mm) and again

medium to large (400 mm).

4.3. Machining power

Due to the lack of a specific standard for test workpieces formachining energy testing, a test workpiece was derived based onthe suggestion of the JSA standard. The limitation of the JSAworkpiece was overcome by scaling the dimensions of theworkpiece matching the capabilities of the machine under study.Dimensions were doubled from small to medium and frommedium to large test piece, respectively. This generated anappropriate level of load that allowed comparing different sizesproperly. The developed test piece included 17 different featuresand incorporated face milling, grooving, pocketing, and drillingoperations (Fig. 2).

Six main features can be distinguished in the design shownabove, which were machined in following order: face milling (1),three large grooves (2), three small grooves (3), X-, Y-, and 458-pockets (4), a trochoidal groove (5), and six holes (6). According tothe previous classification of machine tools, three different sizes oftest pieces were designed (small, medium, and large).

The facing operation was achieved by a face-mill with fiveinserts and a 908 angle. The end-milling operations were machinedwith three sizes of a 4-fluted end-mill. A 2-fluted high precisiondrill with a tip angle of 1408 and a TiN/TiAlN multilayer coating wasused for drilling the six holes. Cold rolled AISI/SAE 1018 steel wasused as the workpiece material. Table 2 summarizes the mainparameters for machining the test piece.

5. Experimental setup

A Yokogawa CW240 clamp-on powermeter was used in a three-phase, single-load setup for measuring the power consumptionwhile performing the procedures of the developed methodology.The maximum sampling frequency of the used device is 10 Hz. Intotal, nine different machine tools were studied including four 3-axis vertical milling machines (Mori Seiki (MS) NVD1500 (24,000

1ardized start-up procedure.

Fig. 1. Spindle and axis movement [13].

Fig. 2. Standard test piece (small size).

Component Operation Time [min]

Power meter Start recording 00:00

Main switch On 00:30

Transformera On 01:00

Panel On 01:30

Door Open/close 03:00

ndby power

Panel Off 09:00

Transformera Off 09:30

Main switch Off 10:00

Power meter Stop recording 10:00

nly true for machine tools that are equipped with this component.

and 40,000 rpm), MS Dura Vertical (DV) 5060, and Haas VF-0), a 4-axis horizontal milling center (MS NH8000), two 5-axis verticalmilling machines (MS NMV1500 and MS NMV5000), a mill-turncenter (MS NT1000), and a CNC lathe (MS NL200SY).

6. Experimental results

The following results on energy consumption were calculatedof at least 40 data points. The resulting averaged values guaranteethe reproducibility of the test procedure for each machine tool.

byree

andlack

forandachp ofum

werh in

an00.ine

tooled),thattool

thene’swery of6 to

nedinein,

ceswas

00

.3

T. Behrendt et al. / CIRP Annals - Manufacturing Technology 61 (2012) 43–46 45

6.1. Standby power

The studied standby power varied significantly across andwithin the three different classes (Fig. 3). Additionally, standbypower increased with the complexity of a machine tool. The smalland complex 5-axis NMV1500, for instance, has a standby powerthat is almost 2 kW above the medium-sized Haas machine. Evenwithin one (medium-sized) class of machine tools the standbypower value varied up to about 3.7 kW.

6.2. Component power

Comparing the obtained spindle data across the three classes ofmachine tool size revealed that the three groups could also beseparated by the power consumption ratio (slope of the graph)(Fig. 4). While small machine tools have a low ratio with a largerange of rotational speeds, large machines, in comparison, have avery high ratio but a rather small range. The maximum powervalue of the nine machines at the maximum spindle speed variedfrom 650 to 2000 W. At similar rotational speeds, the energyconsumption varied significantly across different machines anddepends on the torque provided by the motors. At 5000 rpm, forexample, the power demand ranged from 78 W for the NVD1500(40,000 rpm-version) up to 1265 W for the NH8000.

The spindle speed and power demand for the turning spindlesof the NT1000 and NL2000 follow a linear relationship. Since theworkpiece was mounted directly to the spindle, the motors had toprovide significantly more torque than the milling spindles. Thisresulted in power demands of about 2–9 kW at 5000 rpm.

In addition, the power consumed by the remaining componentssuch as coolant pumps, chip conveyor, tool changer, and drives wasassessed using the components cycle. Sankey diagrams were

created for each machine for visualizing the energy consumedeach component relative to the overall energy consumption. Thdifferent shadings were chosen for separating the energy

easing the analysis of gathered data through colored coding: bfor the components that contribute to standby power, dark graycomponents in run-time mode, and light gray for the spindle

axes with variable energy consumption. The thickness of earrow reflects the percentage on overall energy, shown at the tothe diagram. The spindle power is given for 50% of the maximspindle speed. The axis power is given for the average poconsumed by each axis at a feed rate of 2500 mm/min, whicmost cases is also 50% of the maximum cutting feed rate. Asexample, Fig. 5 shows the component power of the MS NH80

The main power consuming components of all of the nmachine tools were the coolant pumps, spindle, controller,

change system (only using energy when tool change is performand the hydraulics. Fig. 6 shows that the components

accumulate to the constant power demand of the machine

account for up to 41% on total power demand. However, these arecomponents that are always turned on, regardless of the machioperational status. Again, this figure shows that the total poconsumption is highly dependent on both the size and complexitthe machine tool. In this study, total power varied from about 1.19.4 kW.

6.3. Machining power

The contribution of the material removal process is determiby machining the standard test piece on six of the nine machtools. As indicated in Table 3, cycle times (TCycle) of 10:45 m15 min, and 1:16:15 h for the small, medium, and large test piewere designed. The energy (E) used during this process

Table 2Cutting parameters (medium-sized test piece).

Operation # 1 2 3 4 5 6

DOC [mm] 1.0, 2.0, 3.0 20.0

Feed rate [mm/min] 160 200

240

280

400

480

560

400 400 100, 200, 300, 400, 500, 6

Feed/tooth [mm/tooth]/feed/revolution [mm/rev] 0.05 0.05

0.06

0.07

0.05

0.06

0.07

0.05 0.05 0.05, 0.1, 0.15, 0.2, 0.25, 0

Fig. 3. Standby power of studied machine tools. Fig. 5. Component power of the NH8000.

Fig. 4. Spindle power of studied machine tools. Fig. 6. Component power of studied machine tools.

comusedNH8usedmeathe

EAir

and

howNMVcontwhi3.3 kfactoelecby intranmaccurrthe

charmac

6.4.

Tdizecomcertthe

sumcomthe ma desourcomtow

7. C

Eincrto danaldemstanof mtoolproc

Tgrouto tdommacmaccomto toptispinaddr

TableChara

Val

TCy

E [k

qCu

PPe

PF

T. Behrendt et al. / CIRP Annals - Manufacturing Technology 61 (2012) 43–4646

puted by integrating the power data over TCycle. The NVD1500 0.19 kWh for machining the small test piece, while the000 used 9.3 kWh for the large test piece. The energy that is

for the cutting process only (ECut) was determined bysuring at first the energy used for performing an air cut (EAir) ofexact same cycle (i.e. no material was removed). Subtractingfrom E reveals ECut, which is given as a percentage of E (qCut)varies between 2% and 20%. For the complex machines,

ever, this value accounts for only 2–8% (NVD1500, NMV1500,5000, and NH8000). This observation shows the little impact

ributed by the machining process. The peak power (PPeak)le machining is essential for power grid design and varied fromW for the NVD1500 to 55.6 kW for the NH8000. The powerr (PF) is an important measure for the energy efficiency of an

tric system. Most power consumed in industry is contributedductive loads (e.g. electric motors, drives, coolant pumps, and

sformers), representing some of the main components ofhine tools that accumulate to a low power factor. Increasedent flow and voltage drops are caused by low PFs, which reduceelectrical system’s distribution capacity and cause penaltyges for PFs less than 85% (plant level) [17]. Five of the sixhine tools have a power factor of less than 70%, Table 3.

Standardized energy reporting

he standardized data can further be used for creating standar-d energy data sheets. In addition to the main characteristics, theponents contributing to standby power, the spindle power atain rotational speeds, the component power consumption, andenergy data obtained through machining the test piece can bemarized. The Sankey diagrams with colored coding support theprehensive display of energy data. In comparison to categorizing

achine toolsregardingthe useofenergy, thedatasheetsprovidetailed overview of the different energy consumptions and theirces. Publishing the energy consumption data would enforce thepetitiveness among manufacturers and foster the activitiesards a reduction of energy use of their machines.

onclusion

fforts to improve the energy demand of machine toolseasingly rely on effective methods and assessment procedureserive functional data as a mandatory input to subsequentysis and improvement. In addition to measuring the powerand in operational states, this paper proposed the use ofdard assessment procedures to characterize the power demandachine tools. The applicability is validated on diverse machine

s indicating the high potential of formalized assessmentedures as a basis for improvement and exchange of information.he presented three-step methodology for measuring the main

machine tool concept. The required data for evaluating andoptimizing the operational behavior of a machine is generated byexecuting all three steps of the presented methodology. Theproposed methodology is useful for companies wanting to establishsome standard practices until such time as a standard is established.It can also provide useful input to the standards organization withrespect to basic requirements for standard parts (i.e. scalability,accommodating critical elements, ease of implementation, etc.) [15].

Acknowledgments

This work was supported in part by Mori Seiki, the DigitalTechnology Laboratory (DTL), the Machine Tool TechnologyResearch Foundation (MTTRF), Sandvik, and other industrialpartners of the Laboratory for Manufacturing and Sustainability(LMAS). The authors would also like to thank Prof. D. Dornfeld andProf. C. Herrmann for valuable insight and advice.

References

[1] U.S. Energy Information Administration (2011) Annual Energy Review 2010,Report No. DOE/EIA-0384 (2010), http://www.eia.gov/totalenergy/data/annual/.

[2] European Union. (2008) Draft Working Plan of the Ecodesign Directive (2009–2011). http://www.eup-network.de/fileadmin/user_upload/Produktgruppen/Arbeitsplan/DraftWorkingPlan_28Apr08.pdf.

[3] Dahmus JB, Gutowski TG (2004) An Environmental Analysis of Machining.Proceedings of 2004 ASME International Mechanical Engineering Congress andRD&D Expo. Paper IMECE2004-62600.

[4] Gutowski TG, Murphy CF, Allen DT, Bauer DJ, Bras B, Piwonka TS, Sheng PS,Sutherland JW, Thurston DL, Wolff EE (2005) Environmentally Benign Man-ufacturing: Observations from Japan, Europe and the United States. Journal ofCleaner Production 13(1):1–17.

[5] Li W, Kara S (2011) Unit Process Energy Consumption Models for Manufactur-ing Processes. Annals of the CIRP 60(1):37–40.

[6] Li W, Zein A, Kara S, Herrmann C (2011) An Investigation into Fixed EnergyConsumption of Machine Tools. Proceedings of the 18th CIRP InternationalConference on Life Cycle Engineering (Braunschweig, Germany), 268–273.

[7] Mori M, Fujishima M, Inamasu Y, Oda Y (2011) A Study on Energy EfficiencyImprovement for Machine Tools. Annals of the CIRP 60(1):145–148.

[8] Herrmann C, Thiede S, Zein A, Ihlenfeldt S, Blau P (2009) Energy Efficiency ofMachine Tools: Extending the Perspective. 42nd CIRP Conference on Manufac-turing Systems (Grenoble, France).

[9] Diaz N, Helu M, Jarvis A, Tonissen S, Dornfeld D, Schlosser R (2009) Strategiesfor Minimum Energy Operation for Precision Machining. Proceedings of theMachine Tool Technologies Research Foundation 2009 Annual Meeting (Shanghai,China), 47–50.

[10] Diaz N, Redelsheimer E, Dornfeld D (2011) Energy Consumption and ReductionStrategies for Milling Machine Tool Use. Proceedings of the 18th CIRP Interna-tional Conference on Life Cycle Engineering (Braunschweig, Germany), 263–267.

[11] Dornfeld D (2010) Standards for Environmental Performance in Manufactur-ing. Environmental Leader – Environmental & Energy Management News. http://www.environmentalleader.com/2010/09/27/standards-for-environmental-performance-in-manufacturing/.

[12] Kuhrke B, Schrems S, Eisele C, Abele E (2010) Methodology to Assess theEnergy Consumption of Cutting Machine Tools. Proceedings of the 17th CIRPInternational Conference on Life Cycle Engineering (Hefei, China), 76–82.

3cteristics of machining process.

ue Small Medium Large

NVD1500 NMV1500 Haas VF-0 DV5060 NMV5000 NH8000

cle [h] 0:10:45 0:15:00 1:16:15

Wh] 0.1845 0.7142 0.3638 0.5557 1.9323 9.2907

t [%] 3.65 1.85 20.53 14.7 4.25 7.57

ak [W] 3320 20,130 15,610 16,440 36,900 55,600

[%] 98 64 66 69 51 69

ps of power demand can be applied by potential users accordinghe desired level of the detail. First, identifying the highlyinant idle power might be sufficient for sole customers ofhine tools. Optimization strategies that aim to switch thehine in idle mode when not in use can now be evaluated. Second,ponent manufacturers might expand their assessment activitieshe second step, which allows evaluating the effectiveness ofmization efforts regarding specific components such as drives,dles, and coolant pumps. Including the third step mainlyesses machine tool manufacturers for questioning the overall

[13] Japanese Standards Association (2010) Machine Tools—Test Methods forElectric Power Consumption–Part 1: Machining Centres, TS B 0024-1:2010.

[14] Abele E, Kuhrke B, Rothenbucher S (2010) Entwicklungstrends zur Erhohungund Bewertung der Energieeffizienz spanender Werkzeugmaschinen. Proceed-ings of the 1st International Colloquium of the Cluster of Excellence eniPROD(Chemnitz, Germany), 99–120.

[15] International Organization for Standardization. (2009) ISO/NP 14955–Environ-mental Evaluation of Machine Tools. http://www.iso.org/iso/iso_catalogue/cat-alogue_tc/catalogue_detail.htm?csnumber=55294.

[16] Hagemann D (2011) 2nd Stakeholdermeeting Lot 5 EuP—Status of ISO/TC39/WG12. http://www.ecomachinetools.eu/typo/meetings.html?file=tl_files/pdf/Statusreport_ISOWG12_28032011.pdf.

[17] U.S. Department of Energy. (2009) Reducing Power Factor Cost. http://www1.eere.energy.gov/industry/bestpractices/pdfs/mc60405.pdf.