Integrated methodology for the evaluation of the energy- and cost-effectiveness of machine tools

CIRP Journal of Manufacturing Science and Technology 5 (2012) 151–163

Contents lists available at SciVerse ScienceDirect

CIRP Journal of Manufacturing Science and Technology journal homepage: www.elsevier.com/locate/cirpj

Integrated methodology for the evaluation of the energy- and cost-effectiveness of machine tools§ U. Go¨tze a, H.-J. Koriath b, A. Kolesnikov c, R. Lindner a, J. Paetzold c,* a

Management Accounting and Control, Chemnitz University of Technology, Chemnitz, Germany Fraunhofer Institute for Machine Tools and Forming Technology IWU, Chemnitz, Germany c Institute for Machine Tools and Production Processes (IWP), Chemnitz University of Technology, Chemnitz, Germany b

A R T I C L E I N F O

A B S T R A C T

Article history: Available online 2 June 2012

Sustainable machine tools are characterized by less consumption of energy as well as low costs during their life-cycle. In order to design such machine tools an evaluation of the energy-efficiency and the costs incurred is needed. In contrast to this, up to now there has been no fully developed evaluation method relating to both technical and cost efficiency of machine tools. Thus, the paper presents an integrated approach for the evaluation of machine tools consisting of methodological proposals for the measurement of energy consumption, modeling of energy flows and simulative analysis of the energy saving potentials as well as an energy-oriented life-cycle costing concept. Additionally, braking energy storage systems and reactive power compensation are analyzed. Based on the identified energy saving potentials, it is shown that at present only reactive power compensation is economically profitable under given assumptions. ß 2012 CIRP.

Keywords: Eco-design directive Evaluation method Machine tools Energy-efficiency Simulation model Braking energy storage Reactive power compensation Life-cycle costing

1. Introduction In order to ensure sustainable production and environmentalfriendly use of machine tools (MT), customer demands for productivity, flexibility and cost-effectiveness are requested in a global competition. With resources dwindling and energy prices rising, the energy consumption of MTs is an increasingly important factor with respect to ecological as well as economic targets. Efforts to reduce this consumption are reasonable, considering the indication of an energy saving potential of manufacturing systems and processes of 30% in the automotive industry and the production of white goods [1]. Consequently, the European ECODESIGN Directive 2009/125/ EG requests energy saving of new goods, including MTs [2]. Accordingly, considerable effort is made in science and industry to develop MTs or to adjust existing ones with the aim of reducing energy consumption while maintaining performance and costeffectiveness. For this reason, the draft standards ISO 20140 ‘‘Environmental and energy efficiency evaluation method for manufacturing system’’, ISO 22400 ‘‘Key performance indicators for manufacturing operations management’’ [3], CECIMO ‘‘Self-

§ This paper was originally submitted to the Special Issue ‘Energy Efficient Production’. (Vol 4. Issue 2, 2011). PII: S1755-5817(11)X0005-6. * Corresponding author. Tel.: +49 371 531 38282. E-mail address: [email protected] (J. Paetzold).

1755-5817/$ – see front matter ß 2012 CIRP. http://dx.doi.org/10.1016/j.cirpj.2012.04.001

Regulatory Initiative’’ for energy-efficient MTs [4] and ISO 14955 ‘‘Environmental evaluation of machine tools’’ [5] are under development. Additionally research activities by the CIRP CWG EREE [6] and CO2PE! [7] are noted. Moreover, some basic research results are provided by the eniPROD (‘‘Energy-efficient Product and Process Innovation in Production Engineering’’) Cluster of Excellence [8]. This paper presents a novel methodology for the evaluation of the energy- and cost-effectiveness of machine tools which is characterized by the integration of energy flow measurement, predictive energy consumption simulation of drive systems and life-cycle costing. Section 2 gives an overview of the methodology, its pillars and their interfaces. Section 3 provides procedures for the measurement of the energy consumption as a basis for deriving energy-saving priorities and supporting further technical and economic investigations with an energy-related database. Section 4 describes a model of energy flows in drive systems which was developed to enable simulative analyses and applies this model to the evaluation of the energy-saving potentials of braking energy storage systems and reactive power compensation. Afterwards, Section 5 presents an energy-oriented life-cycle costing concept that is used for the economic evaluation of the proposed technical measures using the results of the energetic evaluations. Thus, the methodology allows for a targeted design and decision-making concerning energy-related investments from a technical as well as an economic perspective. Section 6 summarizes the findings and gives an outlook.

152

U. Go¨tze et al. / CIRP Journal of Manufacturing Science and Technology 5 (2012) 151–163

2. Overview of the methodology The methodology proposed is characterized by the combination of different technical and economic evaluation pillars (see Fig. 1). As part of the technical evaluation, the measurement of energy consumption is used to reveal the energy flows and consumptions in components of MTs. The data derived this way can be used in the various steps of the problem-solving procedure; in particular, they allow the identification of the main energy consumers representing starting-points of efforts to improve energy-efficiency. Besides, the data form a basis for the modeling and simulation as well as for the economic evaluation. Using modeling and simulation, design alternatives of identified main consuming components or functions can be analyzed with respect to potential energy savings. This is possible virtually and, thus, in a time- and cost-efficient way. The verification of these models can also be based on the data from the measurements. Modeling and simulation contribute to the problem-solving procedure as well. In particular, they help to identify design alternatives leading to higher energy-efficiency. The design alternatives that are advantageous with respect to energy consumption will then be subjected to an economic evaluation focusing on the life-cycle cost. Therefore, the technical/energy-related data derived from measurements and simulations are used and complemented by economic data, such as prices, e.g. for energy, additional up-front costs, etc. that influence the lifecycle cost of the MT. After the decision based on the different evaluations and the realization of the chosen measures, a verification of the energy-savings is possible using measurements. And, if required, a new cycle of evaluation and problem-solving may start. The technical and economic evaluation of machines is a complex and demanding task due to various design alternatives, influencing factors and long time-periods that have to be considered. Thus, in order to perform the evaluation in a structured and consistent way it is recommended to use a procedure model, as it is shown in Fig. 2. This model consists of two levels: the machine tool-level and the submodel-level. On the ‘‘higher’’ MT-level, the MT as a whole is evaluated with respect to technical/energetic and/ or economic criteria. In a subordinate submodel-level, evaluations can be made referring to MT components or processes within the MT. The results of these evaluations provide a basis for the improvement of specific components and processes and contribute to the evaluation of the whole MT.

At both levels the overall evaluation procedure is subdivided into six steps. In the beginning, the target figure(s) and the basic evaluation method (step 1) as well as the system boundaries and evaluation period(s) (step 2) have to be defined. Afterwards, the MT can be modeled and analyzed (step 3) and its relevant environmental conditions investigated (step 4). At the MT-level this will typically lead to the identification of relevant evaluation subtasks (e.g. as regards the components and functions of the MT) and a turn to the submodel level. In step 5, further relevant data have to be gathered and forecast, until the target figure(s) are calculated and sensitivity analyses are conducted in the last step (referring to the influence of specific variables, uncertainty, etc.). The use of the procedure model for systematic technical as well as economic analyses will be presented in the following. 3. Energy consumption analysis and measurements The aim of energy consumption analysis is to specify the energy flows within the MT, depending on power consumption and operating time. The following approach for energy analysis focuses on energy consumption of MTs in different modes of operation in order to identify main energy consuming components and functions and priorities of their energy-reduction; it is not focused on the optimization of manufacturing processes and process time. The methodology described below is based on the procedure model (Fig. 2) and is exemplarily applied to cutting MTs. 3.1. Step 1 – definition of target figure(s) and basic evaluation method(s) At the first step (M1), energy consumption of a cutting MT is defined as target figure. The main energy sources for this object are electricity and compressed air. The first one dominates the environmental impact of MTs at their entire life-cycle with a share of more than 90% [2]. Therefore the following considerations focus on this kind of energy. Energy within material flows will not be considered. The evaluation is based on the input-throughput-output (ITO) model with energy as input and process energy and energy losses as output (Fig. 3). In general, measurement can be used to record actual energy consumption, while simulation is usable to calculate actual and additionally expected consumption data. The results of measurement and/or simulation are aggregated to energy

Fig. 1. Methodology for the technical and economic evaluation of MTs.

U. Go¨tze et al. / CIRP Journal of Manufacturing Science and Technology 5 (2012) 151–163

153

Fig. 2. Procedure model for technical and economic evaluation of MTs. Modified and adopted from [9].

consumption figures of MTs (and their components, functions) which may be compared and from which performance indicators as power factors and energy efficiency may be derived.

manufacturing operations under industrial production conditions, periods of several minutes up to days or months are needed. 3.3. Step 3 – structural analysis and modeling

3.2. Step 2 – definition of system boundaries and evaluation period(s) At step M2, system boundaries and evaluation periods have to be defined. The scheme in Fig. 3 shows the system boundaries (modified from [10]). On the input side, the infrastructure of the factory supplies the needed energy flows. On the output side, the energy can be divided into process energy and energy losses (see also Fig. 6). In principle, the evaluation period extends over the entire operational phase of the life-cycle of a MT. The average life span for numerically controlled MTs is approximately 10 years [11]. Since data about such periods are currently not available or difficult to acquire, experimental analyses over short and, as far as possible, representative periods of time have to be conducted. They can serve as a basis for the validation of simulations as well for comparisons and predictions of energy consumption. Depending on the

The structural analysis (M3) detects and systematizes the energy-consuming sub-systems and provides a qualitative overview of the energy flows. This abstract energetic model of a MT is a fundamental input for simulation modeling. In a first step, the draft of ISO 14955-1 aims in a first step at an eco-design methodology and, thus describes MT functionality independently from processes, technology and design [5]. It will assign energy consumption to functions and afterwards to modules and components. According to this, the energetic systematization shown as an example in Fig. 4 refers to these aspects. The scheme of the sub-systems is oriented on energy flows and functions. Additionally, it includes the nominal power of modules and components (from machine documentation) as well as measuring points for the data acquisition (step 5) and modeling levels for the simulation (Section 4). The specification of voltage levels is required to

Fig. 3. ITO-model with boundaries for energetic analysis of a MT.

154

U. Go¨tze et al. / CIRP Journal of Manufacturing Science and Technology 5 (2012) 151–163

Fig. 4. Application example for energy flow oriented systematization of sub-systems of a milling center.

choose the right measuring method. The factory infrastructure provides the MT with electric and pneumatic energy (input) and the factory infrastructure and the environment absorb the loss of energy as waste heat (output). The sub-systems inside the MT transform and distribute energy (throughput) in order to enable or support the milling process (output). If partial analysis of these sub-systems are required, for instance for the servo drive unit, the algorithm at Fig. 2 can be applied from step SC1 to SC6. 3.4. Step 4 – determination of environmental scenarios and boundary conditions The components and functions can work (cutting-) processdependently (servo drive unit, coolant) or process-independently (others). This provides the starting point for a delineation of operating states as an important boundary condition (M4). In a simplified way, the operating states can be distinguished between

‘‘Off, Stand-by, Ready for production, Production’’ [10], while the draft of ISO 14955 [5] will differentiate more steps. However, in [12] it is shown that particular operating states enable an approximate forecast for energy consumption. The specific energy requirements and the appropriate dwelling time in these states are central elements required to calculate the total energy consumption during the operational phase of the MT life-cycle. Important parts of the environmental scenario are the typical process parameters. They are difficult to determine for universal machines and for unknown operation conditions. This applies in particular when the entire operational phase of life-cycle is under consideration. Based on typical annual production quantities for various types of manufacturing (mass production, line production, etc.) and the resulting modes of operation, corresponding energy consumption scenarios have to be developed. At the moment, typical manufacturing processes for the analyzed MT have to be defined case-by-case. Especially, the definition of workpieces,

U. Go¨tze et al. / CIRP Journal of Manufacturing Science and Technology 5 (2012) 151–163

155

Fig. 5. Measurement of apparent, active and reactive power and energy consumption during a milling process.

which represent typical manufacturing processes and machining cycles for different machines adequately, is an unsolved problem that will be focused on in parts 3 and 4 of draft ISO 14955 [5]. In the case of the presented MT analysis, a milling process was chosen for the studied machining center: a carbide-tipped cutter with 18 mm diameter (3008 rpm, 170 mm/min feed rate, 2 mm depth and 10 mm width of cut) mills three times a 30 mm long block of steel. 3.5. Step 5 – gathering/forecast of data The data acquisition (M5) includes the metrological determination of the previously defined input and output variables and, if necessary, calculations and simulations which are based on plausible assumptions. The definition of measurement points is based on the pre-established systematization (M4) and is shown in Fig. 4. The specific metering points at the servo drive unit are chosen with respect to the aim of determining the energy consumption of important single components and validating simulation models. The measurement methods have to be adapted to these points. Fig. 5 shows electric power measuring results at the main power supply of the MT for the milling process with low pressure coolant. The active power profile already displays the recovery of braking energy (negative peaks) during the deceleration of the spindle and the axes. The difference between the higher (positive) peaks of acceleration and the lower (negative) ones of deceleration show energy saving potential in these components. Furthermore, the amount of reactive power, 3.3 kVar on average, indicates the potentials for reducing energy by compensation. The measured consumption of compressed air (approximately 1.8 Nm3/h) is not shown in the diagrams and was converted into an additional electrical power consumption with 0.107 kW/ (Nm3/h) (average value from [13]). 3.6. Step 6 – determination of target figure outcomes and sensitivity analyses In the last step (M6), the target figures are determined, input and output variables are compared and the plausibility of the

values is proven. Results of partial analyses are included in the overall balance sheet or can be used for further calculations (e.g. for components). As shown in Fig. 5, the cutting process (milling) needs only a small part of the production time and, consequently, of the energy consumed by the MT. Thus, a power balance for the central operation mode ‘‘production’’ (and not an energy balance) is prepared to eliminate the factor ‘‘time’’ as another influencing variable. Fig. 6 shows the shares of electrical power consumption by different functions. The shaping process requires only 14% of the power at the main power supply, while 40% of the power is needed at the servo drive unit. Therefore, there is a strong priority to take into account measures for the optimization of its components. Potential improvements of the servo drive unit are seen in the recovery of kinetic energy (acceleration/deceleration) and the compensation of reactive power. They will be analyzed in the following. 4. Modeling of energy flows and simulative analysis of the energy saving potentials 4.1. Modeling Modeling and simulative analysis in MT can provide metrologically difficult measurements as well as advice on energy efficiency with relatively little effort, permitting to analyze the electric energy flows for their optimization and to predict efficiency potentials of different measures for energy efficiency [14]. As so far there has been no simulation model that focuses on the optimization of electrical energy cycles in MTs, it is necessary to create one with sufficient accuracy. Therefore Matlab/ Simulink software has been chosen for the modeling and the SimPowerSystems toolbox has been used for the simulations. The simulation model is based on the metrological analysis of energy consumption, described in Section 3, and mechatronic modeling [15]. Since highly dynamic process sequences play an important role for the power input [16], the ability to alter the dynamic properties of the drive unit becomes essential and the corresponding relations were worked out in order to create a model.

156

U. Go¨tze et al. / CIRP Journal of Manufacturing Science and Technology 5 (2012) 151–163

Fig. 6. Electrical power consumption of main functions of a machining center during machining.

The developed drive system model, which is shown in its basic structure, can be divided into three levels (Fig. 7). 4.2. Level A: Individual drives of the drive unit This level includes the individual electric drives of the drive unit in the machining center. For the modeling of the spindle and axis feed drives, SimPowerSystems toolbox elements were used which describe the permanent synchronous motors and linear motors of the axis feed drives as well as the asynchronous spindle and their converters with sufficient accuracy. The mathematical description of the electric motors is based on Ampe`re’s force law and Kirchhoff’s and Newton’s second laws [17]. The modeling of the motor-side converters relies on the mathematical description of conventional semiconductors [18]. The vector and speed controller elements on the basis of PI regulation were used to create the control of the converters.

This level is important for examining the amount of energy which can be recovered from the braking process of the spindle and the axis feed drives at different speeds, thus considering efficiency. The amount of dynamic energy Edyn resulting from the braking process, which could be used for the next acceleration processes of the spindle and the axis feed drives, can be estimated by the following equation [19]: Edyn ¼

1 1 ðM mot ðnmot max  nmot min Þ  hGr Þ  t þ 2  9:55 2  ðF mot ðV mot max  V mot min Þ  hGl Þ  t;

hG ¼ hM  hINV ; hG ¼

1

hM  hINV

;

(1)

for braking

(2)

for acceleration

(3)

Fig. 7. Drive system block model of the machining center in Matlab SimPowerSystems.

U. Go¨tze et al. / CIRP Journal of Manufacturing Science and Technology 5 (2012) 151–163

with Mmot is the max. motor torque when braking or accelerating, nmot max is the max. speed at the start or end of the operation, nmot min is the min. speed at the start or end of the operation, Fmot is the max. motor force during braking or accelerating, Vmot max is the max. velocity at the start or end of the operation, Vmot min is the min. velocity at the start or end of the operation, t is the time of the operation, hGr is the total efficiency of rotating drives, hGl is the total efficiency of linear drives, hG is the overall efficiency usable for rotatory efficiency hGr and linear efficiency hGl, hM is the motor efficiency, hINV is the inverter efficiency.

157

line-to-neutral voltage and current during no-load operation, h is the efficiency of the electrical motor. The model of the above-mentioned electric drives can be parameterized based on the CNC command syntax which supplies the set point and feed-forward control values for the model. Modeling the processing path requires temporarily consistent and given process forces. For this reason, cutting forces were metrologically determined in initial experiments and were used as disturbance variables for the electric drives of the model [15]. 4.5. Braking energy storage system

4.3. Level B: Line-side converter of the drive unit The practical line-side converter was designed with the line reactor and the capacitance in the DC link. The mathematical description of the line-side converter is based on the same description of conventional semiconductors as mentioned in level A. The control-strategy is realized on PI regulation. The Level B is the most important part of the model. On the one hand the line-side converter includes the DC link, which is connected with the energy storage device. On the other hand this converter can compensate the fundamental reactive power component Q1 and the influence of the harmonic reactive power component D in different operating conditions during unstable oscillation, which could be estimated for one phase as: vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi sffiffiffiffiffiffiffiffiffiffiffi!2 u 1 qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi u X 2 t 2 2 Q o p ¼ Q1o p þ Do p ¼ ðU  I1  sin’Þ þ U Iv2 ; (4) v¼2

Qop is the reactive power in one phase, Q1op is the fundamental reactive power component in one phase, Dop is the harmonic reactive power component in one phase, U is the rms (root-meansquare) line-to-neutral voltage, I1 is the rms of the current fundamental component, C1 is the phase angle between U and I1, Iv is the rms current of vth harmonic oscillations, v is the number of the harmonic component. 4.4. Level C: All energy consumers in the machining center (including the regulated drive unit with its spindle and feed drives and the unregulated auxiliary drive systems)

To analyze the braking energy cycles, the power consumption was examined in different drives of the drive unit. Fig. 8 displays almost similar active power and energy consumption of the drive unit at the simulation and the measurement during a part of a milling operation. It shows that the divergence between the simulation and the measurements in terms of energy consumption is less than 0.7%. Consequently, the model allows an exact analysis of the regenerative feedback power, flowing to the DC link and also from the DC link into the network during operating of the servo drives as a generator. The usage of the developed model permits to determine the energy consumption and the reduction of peak power by storing the braking energy. Usually, a part of braking energy is stored in the DC link by electrolyte capacitors. The conventional electrolyte capacitors have a suitable specific power of approx. 1000 kW/kg, but their energy density is less than 0.1 Wh/kg, which is not enough to prevent peak power in machining centers with limited external dimensions [20]. Therefore the usage of double layer capacitors, batteries, or their combination (hybrid batteries) was taken into account. Batteries have a specific power of less than 1 kW/kg and their exclusive usage is unsuitable for the prevention of peak power in machining centers. The implementation of a hybrid battery system which has a high specific power by including double layer capacitors and a good outcome for preventing the peak power can increase specific power. These energy storage systems are applied and verified in the branch of renewable energy [21]. However, hybrid batteries in machining centers require a higher

Precise comprehension of the interaction between the auxiliary drive systems and the drive unit allows the analysis of reactive power consumption in the whole machine and its compensation. The auxiliary drives are the group of components with the most unregulated asynchronous motors (pumps) which consume not only active power, but also the fundamental reactive power component Q1. The mathematical description of the asynchronous motors is based on the same laws as mentioned in level A. There is a constant part of reactive power during the no-load operation Qnl and a load dependent part DQld, which could be estimated by the following equations: pffiffiffi Q nl ¼ 3  I  U ll  sin ’; (5)

DQ ld ¼

PN

h

tg ’  2

pffiffiffi 3  I  U ll  sin ’;

Q 1 ¼ Q nl þ b  DQ ld ;

b¼

P ; PN

(6) (7) (8)

Qnl is the reactive power during no-load operation, DQld is the loaddependent reactive power, I is the rms current during no-load operation, Ull is the nominal line-to-line voltage, PN is the nominal power, P is the actual active power, w is the phase angle between

Fig. 8. Electric energy and power during a part of a milling operation.

U. Go¨tze et al. / CIRP Journal of Manufacturing Science and Technology 5 (2012) 151–163

158

investment outlay than double layer capacitors for which a specific power of approx. 10 kW/kg is enough to prevent the peak power [22]. These capacitors have an overall efficiency of 85–98%, depending on the dimension and the type of the storage system. The amount of stored electrical energy in supercapacitors can be calculated on basis of the considerations below. The overall capacitance and voltage can be calculated as: n 1 X 1 ¼ ; C C i¼1 n

U ov ¼

n X Un ;

(9)

i¼1

C is the overall capacity, n is the number of capacitors, Cn is the individual capacity, Uov is the overall voltage, Un is the individual voltage. In order to prevent overvoltages at the individual capacitors due to the voltage change 280 capacitors with 2.5 V are used to permit a maximum theoretical voltage of 700 V in the DC link: U ov ¼

280 X 2:5 ¼ 700 V:

(10)

i¼1

An overall capacity is estimated by (9) with an individual capacity of 600 F: C¼

280 X i¼1

1 600 F

4.6. Reactive power compensation

!1 ¼ 2:1429 F

(11)

It must be considered that the inclusion of a passive energy storage system with a direct connection to the DC link uses up to 30% of the total stored energy. For example, a voltage change of 20 V allows a dynamic storage of: Es ¼

The model allows to analyze the braking energy recovery with and without supercapacitors when operating the servo drives as a generator. Thus, it enables the determination of the saving potentials through the optimization of the braking energy cycles. The energy storage system based on double-layer capacitors should be calculated and compared concerning the energy saving potential which could be achieved by reducing the energy consumption, decreasing the peak power, increasing the robustness to network voltage dips and, consequently, lowering the dimensions of the line-side converter and the cross sections in input conductors. The main advantages of using braking energy are the direct storage of electrical energy in the DC link to reduce the amount of energy drawn from the network and the reduction of peak power by 85.6% during the acceleration of the motor spindle up to 3000 rpm. Consequently, the implementation of the energy storage system into the DC link solves the problem of an optimal dimension for the line-side converter and gives the possibility to choose a converter with lower input power. In this case a lowering from 36 kW to 16 kW could be achieved and, hence the nominal power losses would be decreased by 0.265 kW.

2 C  DUov ¼ 428:58 W s; 2

(12)

Es is the stored energy, DUov is the overall voltage change. Fig. 9 shows the simulation results of energy storage system tests in the DC link with and without supercapacitors. The acceleration of the spindle up to 3000 rpm using the energy storage system allows decreasing the energy consumption required for acceleration by approximately 77%. Besides, peak power is reduced by 85.6%, permitting smaller dimensions of the line-side converter and, thus, the reduction of power losses. The dimension and potential benefits of using a braking energy storage system were examined in the drive unit by the simulation.

The most important reason for the compensation of reactive power is the reduction of its fundamental component Q1 [23]. The electrical auxiliary systems include primarily asynchronous motors with a relatively high proportion of the fundamental reactive power which leads to an increase of apparent power S (see Fig. 10 on the left). Moreover, the harmonic reactive power component D caused by semiconductors also influences the apparent power S in different operating conditions, causes additional losses and worsens the overall power factor of the MT l [24], which can be estimated as: S2 ¼ P2 þ Q12 þ D2 ; P S

P

l ¼ ¼ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ; P 2 þ Q12 þ D2

(13) (14)

P is the overall active power. A possible solution for the problem is the compensation of the overall reactive power Q applying the line-side converter to the respective quadrants of operation (Fig. 10 on the right). The line-side converter with the line reactor permits the fundamental and harmonic reactive power compensation. In this case, the amount of the reactive power for the compensation must be defined by the following equation: Q¼

3  US ðU  cos d  U S Þ; vL C

(15)

US is the line-side voltage, UC is the CEMF (Counter Electro Motive Force) voltage, L is the inductance of the line current smoothing inductor, v is the angular frequency, d is the phase angle between US and UC.

Fig. 9. Electric power and energy during acceleration of the spindle up to 3000 rpm.

Fig. 10. Vector diagram and quadrants of the power in feedback-capable line-side converters.

U. Go¨tze et al. / CIRP Journal of Manufacturing Science and Technology 5 (2012) 151–163

159

Fig. 11. Simulative reactive power profiles.

The simulation results of the fundamental reactive power compensation are displayed in Fig. 11. The graphs indicate the full compensation of the overall reactive power Q into the considered range. Additional advantages of the fundamental and harmonic reactive power compensation are:     

decrease of apparent power input into the machining center, avoidance of central reactive power compensation, smaller dimensions of the electrical elements in MTs, improvement of the power quality in the low voltage network, reduction of active energy losses caused by reactive power to nearly zero.

Summarising two proposals were defined due to the metrological analysis from Section 3 and simulatively estimated with respect to energy efficiency improvement. The quantitative results of the simulation for the reactive power compensation of auxiliary drives such as process coolant pump and the braking energy storage during spindle drive accelerations will be used for the economic evaluation in the following chapter. 5. Life-cycle costing for long-term economic evaluation of measures Energy evaluation helps to identify technically and ecologically weak points of MTs. However, an energy-optimized machine does not necessarily meet the criterion of cost-effectiveness. Therefore MTs should not only be evaluated from a technical but also from an economic perspective. This evaluation should cover at least most of the periods of the entire life-cycle. Life-cycle costing (LCC) is predestined to fulfill this purpose and, thus, a concept for energy-oriented LCC will be presented in the following. This will help to further improve the theory and practice of evaluation and design of (energy) efficient MTs (for existing approaches considering LCC for MTs in general, without focus on energy see [25,26]). The utilization of the concept will be demonstrated by two examples with the expected energy consumption included, derived from power measurements (Section 3) and simulations (Section 4) and based on assumptions on operating time during the life cycle. MT differences in energy efficiency normally cause varying costs in several phases of the life-cycle. This concerns the up-front and follow-up costs (mainly in the phases before and during utilization) which are often interdependent. For this reason, an orientation toward up-front or acquisition costs could lead to a decision that is not economic in the long run (see the example in

Fig. 12). After a few years, the higher up-front costs may become overcompensated by the cumulated differences between the follow-up costs during utilization of MT A, respectively, MT B. Within the LCC, models are formulated and evaluated in order to draw conclusions about the MT design. For analyzing LCCmodels it is recommended to use a procedure model as shown in Fig. 2 (see more procedure models for economic evaluation in [28,29]; suggestions for modeling single elements within those economic calculations are presented by [30–32]). In the following, the steps of the procedure model are adapted for the economic evaluation. At the MT-level, alternative machines are evaluated with regard to the life-cycle costs as a whole. Thereby energy costs as a type of these costs play an important role and should be focused on. The life-cycle costs are determined by diverse influencing variables [33,34], including a series of decisions that have to be taken in the different phases of the life-cycle and that are related to the design of the MTs (about components, functionalities, equipment for improving the energy efficiency, etc.) and their utilization (regarding manufacturing methods, tools and equipment, automation, etc.). Using the submodel-level evaluation subtasks such as modeling and analyzing of partial decisions as well as forecasts of the outcomes and impacts of influencing variables can be performed. The results of these calculations provide a data basis for the MT-level which is derived in a structured and consistent way. The presented examples of braking energy storage and the line-side converter with reactive power compensation can be attributed to this submodel-level. Thus, the steps of the procedure model are described with respect to this level. In S1, it is assumed that the emphasis is placed on the reduction of life-cycle costs of MTs by the considered measures (without worsening the specific functionality, ecological compatibility, etc.). As MTs and their design alternatives are characterized by a long lifespan [35], dynamic investment appraisal methods like the net present value method should be used. In order to calculate the corresponding target figures considering long-term economic effects (including interest and compound interest) adequately payments instead of costs should be used. A time horizon of 10 years is considered, which is divided into periods (years) (S2). In S3, the braking energy storage and the line-side converter as (sub-) components of the MT are modeled in order to reveal their energetic properties and other relevant characteristics (see Section 4). Afterwards, scenarios are formulated for the technical, legal and economic factors in the environment of the MT components that are expected to influence the life-cycle costs (S4) (for such factors see [28,36]). This includes utilization scenarios describing the amount of workpieces to be manufactured, the operating times

160

U. Go¨tze et al. / CIRP Journal of Manufacturing Science and Technology 5 (2012) 151–163

Fig. 12. LCC and purchasing decisions for MTs. Modified and adopted from [27].

and modes and the ‘‘surrounding’’ manufacturing, logistics and maintenance processes as well as the prognosis of external factors such as prices, e.g. for energy. Based on this, the economically relevant impacts such as (reactive) energy consumptions or energy recovery losses are determined in S5 by measurements and/or analytical and simulative calculations (see [36] and Sections 3 and 4). This enables a determination of the resulting energy and other costs (or cost differences) by valuing the consumptions with energy and other prices. In order to get highly significant results, all relevant costs should be taken into account in a sophisticated way – differentiating between types of costs, the costs of components and processes and of periods, as shown by the example in Fig. 13 (see also [28,30,32]). In the case of the braking energy storage, the impacts on the energy costs have to be identified (further costs of the utilization phase such as maintenance costs are ignored in the following). A major effect of reducing energy consumption and, thus, energy cost can be seen in the decrease of energy losses caused by storing and reusing of braking energy. This effect depends on the number of braking applications (e.g. tool changing). A further effect results

Fig. 13. Cost elements of MTs. Modified and adopted from [31].

from the buffering of braking energy – the line-side converter now requires less nominal power. This leads to lower energy consumptions and, thus, to more savings in energy costs. These reductions of costs overall depend on the agreed energy consumption charge for electrical energy that is purchased from the system operator. Furthermore, they depend on operating time of the MT. Additionally, load balancing effects are noted which lead to lower power consumption and a smaller yearly demand charge. This second part of electricity tariff (besides the energy consumption charge for each purchased kWh) is usually calculated on the basis of the highest arithmetic average relating to a quarter-hour. As already mentioned, assumptions concerning the operation modus have to be made. Here, based on an expected manufacturing program, it is assumed that the operating hours of the MT within the period of 1 year amount to 6000 and that 150 brake applications are carried out per hour. In addition, based on measurement and simulation, every brake application shall circumvent a loss of approx. 570 Ws (equivalent to 0.00016 kWh) in energy recuperation (assuming a spindle speed of 3000 rpm, see Section 4, Fig. 9). This leads to a saving in annual energy consumption of 144 kWh under

U. Go¨tze et al. / CIRP Journal of Manufacturing Science and Technology 5 (2012) 151–163

the simplified assumption that only the operation mode of the MT is considered. Caused by the difference of 0.265 kW per operating hour between the power losses of the modules (see Section 4.2), 1590 kWh can be saved per year. In sum, this results in a decrease of energy consumption by 1734 kWh. On the one hand, this permits an ecologically advantageous reduction of greenhouse gas (GHG) emissions. The relation between the GHG emissions and electricity in Germany currently amounts to approx. 0.563 kg/kWh on average [37]. Thus, on the basis of (19), the reduction of the GHG emissions for the machining center with the energy storage system can be estimated at 976.24 kg per year (without considering the GHG emissions of the measure itself). On the other hand, the economic consequences have to be taken into account. Regarding the present situation, it is assumed that a charge for the emissions is not raised and, thus there is no corresponding saving potential. Then, a saving potential primarily results from the decreased need of energy that has to be purchased: assuming an average energy consumption charge of 0.10 s/kWh, the energy costs can be reduced by the amount of 173.40 s in the first year of usage. Further economic effects can be seen in the reduction of peak power. A lower power consumption of approx. 22 kW at an acceleration time of 0.20 s can be derived from Fig. 9. Extrapolated to 15 min (the amount of demand charge is calculated based on the highest power consumption which was measured within 15 min), this leads to a difference in power consumption of 0.1850 kW referring to 150 brake applications. To include the aspect that often several MTs are combined in a production system a simultaneity factor of 0.25 is considered. Thus, a yearly power consumption of 0.0463 kW results leading to an additional saving of energy cost in the amount of 4.63 s under the assumption of a demand charge of 100 s/kW. In sum, 178.03 s can be saved yearly (see formula (16) and (17)).

DEC B ¼ ðaB  DER þ DPN Þ  h  cC þ DPC  cD

(16)

DECB is the saving in energy costs using a braking energy storage, aB is the amount of break applications, DER is the saving in energy recovery per break application, DPN is the saving in nominal power of the modules, h is the operating hours per period, cC is the average energy consumption charge, DPC is the saving in power consumption per year, cD is the demand power charge.

DEC B ¼ ð150  0:00016 þ 0:265Þ  6000  0:10 þ 0:0463  100 ¼ 173:40 þ 4:63 ¼ 178:03s

(17)

The energy costs of the line-side converter with reactive power compensation were analyzed as well. In the initial case (conventional line-side converter), it was assumed that reactive power of 3.3 kVar (see Section 3, Fig. 5) incurs, which has to be paid additionally. Energy providers will often raise a charge if the power factor (agreed between customer and network operator) has not been adhered to. It is assumed that this is accounted once a year and that 0.01 s/kVarh have to be paid for each unit of excess reactive energy.

DEC L ¼ DQ  h  cR

(18)

DECL is the saving in energy costs using a line-side converter with reactive power compensation, DQ is the saving in reactive power, h is the operating hours per period, cR is the charge for one unit of reactive power.

DEC L ¼ 3; 3  6000  0:01 ¼ 198:00s

(19)

Thus, MTs reactive power costs as a part of the energy costs are yearly 198 s (without any rise in price) in 6000 operating hours

161

(see formula (18) and (19)). These could now be saved using the line-side converter with reactive power compensation. Besides that, an annual reduction of GHG emissions of approx. 16.85 kg can be anticipated for the factory facility (therewith outside the system boundaries). This ecologically positive effect is caused by the decrease of active power losses for the low voltage network by the compensation of 3.3 kVar reactive power. The cost savings mentioned above (formula (17) and (19)) should be diminished by the additional up-front costs (S6). The result is the reduction of the life-cycle costs (referring to t = 0, ‘‘t’’ represents specific points in time) caused by the measures under consideration. This result is expressed as net present value (NPV) and represents the criterion for the profitability. If the NPV is greater than zero, the measure is advantageous. In the simplified example of the braking energy storage, the upfront costs of the energy storage and the converter in the beforeutilization phase (in t = 0) as well as the (reduced) energy costs in the utilization phase (in t = 1, . . ., 10) have to be taken into account (there is no strong reason for considering cost differences in the after-utilization phase). Besides this, the new line-side converter causes lower up-front costs (1000 s). Additionally, the reduced requirements for the infrastructure caused by the leveled power peaks lead to lower up-front (and also lower operating) costs for cables, transformers, fuses, etc. To simplify matters, this aspect is ignored here. Assuming up-front costs for the energy storage of 7000 s (dimensioned for storing the whole braking energy), current costs incurring at the end of each period, a discount rate of 10% and an expected yearly growth in energy prices of 3% (beginning in the first year), the NPV of the storage can be derived (formula (20) and (21)). NPV B ¼ ðC 0B þ DC 0C Þ þ

10 X

ð1 þ gÞt

t¼1

ð1 þ dÞ

DEC B 

t

(20)

NPVB is the net present value of the investment in the braking energy storage, C0B is the up-front costs of the energy storage, DC0C is the savings in the up-front costs of the line-side converter, g is the growth in (energy) prices, d is the discount rate. NPV B ¼ ð7000 þ 1000Þ þ

10 X 178:03  1:03t  1:1t t¼1

(21)

¼ 4738:00s

Because of the negative NPV, the investment in the braking energy storage system has to be considered as not advantageous. Although improving the energy efficiency, it would increase the life-cycle costs of the MT by 4738 s. In contrast, the investment in the (improved) line-side converter with reactive power compensation (up-front costs of 200 s) is advantageous because of a NPV of 1204 s (formula (22) and (23)). The higher up-front costs of the converter are nearly amortized by the energy cost savings of 1 year. NPV L ¼ C 0L þ

10 X

ð1 þ gÞt

t¼1

ð1 þ dÞ

DEC L 

t

(22)

NPVL is the net present value of the investment in the line-side converter with reactive power compensation, C0L is the up-front costs of the line-side converter with reactive power compensation. NPV L ¼ 200 þ

10 X 198  1:03t  1:1t ¼ 1204:00s

(23)

t¼1

However, these results depend on the data used, which will typically be uncertain in particular with regard to future trends (see [31,38]). Therefore, in general and especially in the case of the storage further investigations in S6 are recommended. In particular, these are

162

U. Go¨tze et al. / CIRP Journal of Manufacturing Science and Technology 5 (2012) 151–163

Fig. 14. Sensitivity analysis of the braking energy storage.

sensitivity analyses in the form of if-then calculations and/or the calculation of critical values in order to identify relevant model parameters and their maximum allowable or necessary deviation from the initial values and, thus, to enable selective countermeasures. Fig. 14 shows the course of NPV of the braking energy storage in dependence of input values with 100% as the initial point and taking restrictions concerning some input variables into account (the number of brake applications is limited to 250, the operating hours to 8760 and the simultaneity factor to 1). It can be seen that the up-front costs and the average consumption charges are the most critical factors. If the up-front costs drop to approx. 32% (2262 s) and the average energy consumption charge rises to approx. 500% (0.49 s), cost-efficiency could be reached. Due to experience curve effects in matters of manufacturing, the corresponding price for purchasing the energy storage is expected to fall in the future, the consumption charges will probably rise. Thus, the profitability might change. 6. Summary and outlook The proposed methodology for the energy-oriented evaluation of MTs is characterized by the integration of different technical approaches and an economic appraisal: a uniform procedure model is used and the interfaces between measurement, simulation and economic analysis are defined. Additionally, some methodological pieces of advice for the measurement, simulation and economic evaluation are given. This includes the design of models which can be used for the simulation of innovative energyefficient drive system variants in order to calculate energy saving potentials of braking energy storage systems and reactive power compensation. These energy saving measures are technically and ecologically advantageous, while only reactive power compensation is economically profitable at the moment (with respect to lifecycle costs). Further research should, on the one hand, concentrate on the elaboration and validation of each pillar of the methodology, including the extension of simulation models to other components and functions of MTs, the analysis of ‘‘non-production’’ operating states or the evaluation of a system of connected MTs. On the other hand, a larger data base should be created, particularly in terms of the variety of relevant influencing factors (including operating states) and the energy consumptions depending on them. This would not only contribute to the evaluation of single MTs but also support the establishment of energy performance criteria as it is intended by the initiatives mentioned in the introduction.

Acknowledgments The Cluster of Excellence ‘‘Energy-Efficient Product and Process Innovation in Production Engineering’’ (eniPRODß) is funded by the European Union (European Regional Development Fund) and the Free State of Saxony. References [1] Fraunhofer Gesellschaft. 2008, Energieeffizienz in der Produktion – Untersuchung zum Handlungs-und Forschungsbedarf, Federal Ministry of Education and Research, Chemnitz. pp. 12–13. [2] DIRECTIVE 2009/125/EC of the European Parliament and of the Council of 21 October 2009 establishing a framework for the setting of ecodesign requirements for energy-related products, Official Journal of the European Union, L 285:10. [3] Fukuda, Y., Patzke, R., 2010, Standardization of Key Performance Indicator for Manufacturing Execution System, Proc. SICE Annual Conference 2010 (Taipei, Taiwan), pp. 263–265. [4] CECIMO, 2009, Concept Description for CECIMO’s Self-Regulatory Initiative (SRI) for the Sector Specific Implementation of the Directive 2005/32/EC.(EuP Directive). [5] Hagemann, D., 2011, Status of ISO/TC39/WG12, 2nd Stakeholdermeeting Lot 5 EuP, Frankfurt a.M., . [online] http://www.ecomachinetools.eu/typo/ meetings.html (accessed 25.05.11). [6] CIRP Collaborative Working Group EREE: Energy and Resource Efficiency & Effectiveness: [online] http://www.cirp-eree.iwf.tu-bs.de/ (accessed 25.05.11). [7] Cooperative Effort on Process Emissions in Manufacturing CO2PE!: [online] http://www.mech.kuleuven.be/co2pe!/index.php (accessed 25.05.11). [8] Chemnitz University of Technology, Fraunhofer IWU. 2010, Energy-Efficient Product and Process Innovation in Production Engineering, in Neugebauer R, (Ed.). in: Proceedings of the 1st International Colloquium of the Cluster of Excellence eniPROD. [9] Go¨tze, U., Schmidt, A., Weber, T., 2010, Vorgehensmodell zur Abbildung und Analyse des Lebenszykluserfolgs von Werkstoffen – Konzeption und beispielhafte Veranschaulichung, MWT, 41, vol. 6, pp. 464–475. [10] Neugebauer, R., Frieß, U., Paetzold, J., Wabner, M., Richter, M., 2010, Approach for the Development of Energy-Efficient Machine Tools, XXI International CIRP Sponsorship Conference on Supervising and Diagnostics of Machining Systems in Karpacz (Poland). [11] Hagemann, D., 2010, Die energieeffiziente Werkzeugmaschine – Rahmenbedingungen und Lo¨sungen fu¨r eine nachhaltige Fertigungstechnik, Symposium ‘‘Die energieeffiziente Werkzeugmaschine’’ (Du¨sseldorf, Germany). [12] Dietmair, A., Verl, A., 2009, Energy consumption forecasting and optimisation for tool machines, Modern Machinery (MM) Science Journal Prague, 1. Paper 4. [13] European Commission, Joint Research Centre, 2008, Institute for Prospective Technical Studies: Reference Document on Best Available Techniques for Energy Efficiency, Sevilla. [14] Dietmair, A., Verl, A., 2008, A Generic Energy Consumption Model for Decision Making and Energy Efficiency Optimization in Manufacturing, Flexible Automation and Intelligent Manufacturing, FAIM2008, Sko¨vde, Sweden. [15] Zirn, O., 2008, Machine Tool Analysis – Modelling, Simulation and Control of Machine Tool Manipulators, ETH Zu¨rich, Institute of Machine Tools and Manufacturing, Zu¨rich. [16] Avram, O.I., Xirouchakis, P., 2011, Evaluating the Use Phase Energy Requirements of a Machine Tool System, Journal of Cleaner Production, 19:699–711.

U. Go¨tze et al. / CIRP Journal of Manufacturing Science and Technology 5 (2012) 151–163 [17] Bose, B.K., 2002, Modern Power Electronics and AC Drives, Prentice Hall PTR, Upper Saddle River, NJ. [18] Mohan, N., Undeland, T.M., Robbins, W.P., 2003, Power Electronics, John Wiley & Sons, USA. [19] Configuration Manual – SIMODRIVE 611 Digital Drive Converters, 2008 [online] available: http://www.automation.siemens.com/doconweb/content.asp? item=12410&cd=sinumerik_simodrive_04_2010_e&scope=all. [20] Beres, T., Olejar, M., Dudrik, J., 2010, Bi-Directional DC/DC Converter for Hybrid Battery, in: Proceedings of the 14th International Power Electronics and Motion Control Conference EPE-PEMC-2010 (Ohrid, Macedonia, T9), pp. 78– 81. [21] Tankari, M.A., Camara, M.B., Dakyo, B., Nichita, C., 2010, Efficient Management of Energy Transfer in Wind Diesel Hybrid System-Battery and Ultracapacitors, in: Proceedings of the 14th International Power Electronics and Motion Control Conference EPE-PEMC-2010 (Ohrid, Macedonia, T12), pp. 108–113. [22] Diaz, N., Helu, M., Jarvis, A., To¨nissen, S., Dornfeld, D., Schlosser, R., 2009, Strategies for minimum energy operation for precision machining, The Proceedings of Machine Tool Technologies Research Foundation (MTTRF), Annual Meeting. [23] Neugebauer, R., Kolesnikov, A., Richter, M., Paetzold, J., 2010, Improvement of the power factor correction in machine tools, in: Proceedings of the 14th International Power Electronics and Motion Control Conference EPE-PEMC2010 (Ohrid, Macedonia, T13), pp. 11–16. [24] Dugan, R.C., McGranaghan, M.F., Santoso, S., Beaty, H.W., 2003, Electrical Power Systems Quality, The McGraw-Hill Companies, USA. [25] Osten-Sacken, D.v.d., 1999, Lebenslauforientierte, ganzheitliche Erfolgsrechnung fu¨r Werkzeugmaschinen, Dissertation, Universita¨t Stuttgart, FraunhoferInstitut Produktionstechnik und Automatisierung, pp. 57 ff. [26] Schweiger S, (Ed.), 2009, Lebenszykluskosten optimieren: Paradigmenwechsel fu¨r Anbieter und Nutzer von Investitionsgu¨tern. Wiesbaden.

163

[27] Denkana, B., Harms, A., Jacobsen, J., Mo¨hring, H.-C., Lange, D., Noske, H., 2006, Life-Cycle Oriented Development of Machine Tools, Paper of 13th CIRP International Conference on Life Cycle Engineering (Leuven, Belgium). [28] Abele, E., Dervisopoulos, M., Kuhrke, B., 2009, Bedeutung und Anwendung von Lebenszyklusanalysen bei Werkzeugmaschinen, in Schweiger S, (Ed.). Lebenszykluskosten optimieren: Paradigmenwechsel fu¨r Anbieter und Nutzer von Investitionsgu¨tern. Wiesbaden, pp. 51–80. [29] Blanchard, B.S., 1978, Design and Manage to Life Cycle Cost, M/A Press, Portland (Or.). [30] VDI 2884, 2005, Purchase, Operating and Maintenance of Production Equipment using Life Cycle Costing (LCC). [31] IEC 60300-3-3, 2004, Dependability Management – Part 3-3: Application Guide – Life Cycle Costing. [32] VDMA 34160, 2006, Forecasting Model for Lifecycle Costs of Machines and Plants. [33] Niemann, J., Tichkiewitch, S., Westka¨mper, E. (Eds.), 2009, Design of Sustainable Product Life Cycles, Springer Press, Berlin, Heidelberg. [34] Ehrlenspiel, K., Kiewert, A., Lindemann, U., Hundal, M.S. (Eds.), 2007, CostEfficient Design, Springer Press, Berlin, Heidelberg. [35] Go¨tze, U., Northcott, D., Schuster, P., 2008, Investment Appraisal: Methods and Models, Springer Press, Berlin, Heidelberg. [36] Go¨tze, U., Koriath, H.-J., Kolesnikov, A., Lindner, R., Paetzold, J., Scheffler, C., 2010, Energetische Bilanzierung und Bewertung von Werkzeugmaschinen, in Neugebauer R, (Ed.). Energy-Efficient Product and Process Innovation in Production Engineering, Proceedings of the 1st International Colloquium of the Cluster of Excellence eniPROD pp. 157–184. [37] Umweltbundesamt, March 2011 [online] available in German: http://www. umweltbundesamt.de/energie/archiv/co2-strommix.pdf. [38] Wu¨bbenhorst, K.L., 1986, Life Cycle Costing for Construction Projects, Long Range Planning, vol. 9, no. 4, pp. 87-97.