- Email: [email protected]
Exploring the effect of un-deformed chip parameters on energy consumption for energy efficiency improvement in the milling
Available online at www.sciencedirect.com
ScienceDirect
Available online atonline www.sciencedirect.com Available at www.sciencedirect.com
ScienceDirect ScienceDirect
Procedia CIRP 00 (2018) 000–000
www.elsevier.com/locate/procedia
ProcediaProcedia CIRP 00CIRP (2017) 72000–000 (2018) 1380–1385 www.elsevier.com/locate/procedia
51st CIRP Conference on Manufacturing Systems
Exploring the of un-deformed parameters 28th effect CIRP Design Conference, Maychip 2018, Nantes, France on energy consumption for energy efficiency improvement in the milling A new methodology toYongbing analyze the physical architecture of Wang , Lifunctional li *, Li Lingling and Wei Cai , College of engineering and technology of southwest university, Chongqing,400715, China existing products for an assembly oriented product family identification State Key Laboratory of Mechanical Transmission, Chongqing University, Chongqing, 400030, China a
a
a
b
a
b
Paul Stief *, Jean-Yves Dantan, Alain Etienne, Ali Siadat
* Corresponding author. Tel.: +86-150-2309-2936; fax: +86 023 98251265. E-mail address: [email protected]
École Nationale Supérieure d’Arts et Métiers, Arts et Métiers ParisTech, LCFC EA 4495, 4 Rue Augustin Fresnel, Metz 57078, France
* Corresponding author. Tel.: +33 3 87 37 54 30; E-mail address: [email protected]
Abstract In order to study the influence of milling energy consumption and un-deformed chip parameters, the energy consumption model is Abstract constructed and the validity of the model is analyzed. Firstly, the relationship between un-deformed chip parameters and cutting
is analyzed. Then RSM andtowards centralmore composite is used to get the model. Bytousing the multiple the regression Inparameters today’s business environment, the trend productdesign varietymethod and customization is unbroken. Due this development, need of agile and reconfigurable production to copesurface with various and and product families. To design and optimize production analysis on the experimental datasystems fitting,emerged the response modelproducts of milling cutting parameters in energy consumption is systems as well as toby choose optimalanalysis, product matches, product are needed. of the known methods to obtained. Finally, usingthe variance the model showsanalysis that themethods experimental fittingIndeed, modelmost significantly higher, at theaim same analyze a product one productreduction family on of theun-deformed physical level.chip Different producton families, however, may differ of largely in terms the number and time through theordimension parameters the energy consumption milling planeofwere analyzed. nature of components. This fact impedes an efficient comparison and choice of appropriate product family combinations for the production system. newAuthors. methodology is proposed to analyze © 2018AThe Published by Elsevier B.V. existing products in view of their functional and physical architecture. The aim is to cluster these productsunder in new assembly oriented product families for the optimization existing assembly lines and the creation of future reconfigurable Peer-review responsibility of the scientific committee of the 51st CIRPofConference on Manufacturing Systems. assembly systems. Based on Datum Flow Chain, the physical structure of the products is analyzed. Functional subassemblies are identified, and a Keywords: functional analysis performed. Moreover, a hybrid functionalchip and parameters; physical architecture graph (HyFPAG) is Analysis the outputofwhich depicts the Energyisconsumption of milling; Un-deformed Central composite design; variance; similarity between product families by providing design support to both, production system planners and product designers. An illustrative Dimension reduction example of a nail-clipper is used to explain the proposed methodology. An industrial case study on two product families of steering columns of thyssenkrupp Presta France is then carried out to give a first industrial evaluation of the proposed approach.potential [6][7]. In the process of considerable energy-saving 1. Introduction © 2017 The Authors. Published by Elsevier B.V. manufacture of machine tool structure design, product design Peer-review under responsibility of the scientific committee of the 28th CIRP Conference 2018.optimization, it can effectively reduce andDesign machine parameters
CNC processing, as the main production mode of manufacturing, widely used in identification the machine process of Keywords: Assembly;isDesign method; Family industrial products. However, as the rise in energy prices and awareness of environmental protection, the reduction of electricity consumption and CO2 emission become the driving 1.force Introduction for optimizing energy demand in manufacturing industry [1][2]. The U.S. Energy Information Administration (EIA) Due toan the fast development in2017, the and domain of published International energy outlook the global communication and an ongoing trend of digitization and energy-related carbon dioxide (CO2)emissions are expected digitalization, enterprises facing important billion to rise from manufacturing 33 billion metric tons inare 2015 to 39 challenges in today’s market environments: a continuing metric tons in 2040[3]. Meanwhile, the china government has tendency towards reduction oftoproduct development and also made a long-term policy reducing the carbon times emissions shortened product lifecycles. In addition, there is an increasing [4]. The relative standard they have set is to reduce emissions demand of customization, at the same a global of a particulate matter being by 40%by 2040.time It in forces the competition with competitors all over the world. This manufacturing industry to face the development oftrend, high which is inducing the and development from macro to [5][3]. micro efficiency, high speed low energy consumption markets, results in diminished lot sizes due to augmenting With a wide distribution and large amount of energy product varieties to low-volume production) [1]. consumption at (high-volume a low efficiency, machining process have To cope with this augmenting variety as well as to be able to identify possible optimization potentials in the existing 2212-8271 ©system, 2018 TheitAuthors. Published Elsevier B.V. knowledge production is important tobyhave a precise
the energy consumption, product manufacturing process is the effective way to realize the green of manufacturing high efficiency, low energy consumption and sustainable development [8][13]. For better analysis and evaluate the energy consumption of characteristics the production process ofand/or NC of the product range and manufactured machininginsystem, it is necessary to model and challenge analyse the assembled this system. In this context, the main in energy consumption of CNC machining under different modelling and analysis is now not onlysystem to cope with single technological parameters and or establish the relationship products, a limited product range existing product families, between parameters and products process to energy but also to beprocess able to analyze and to compare define consumption[9][11]. new product families. It can be observed that classical existing In the process NC machining, un-deformed directly product families areofregrouped in function of clientschip or features. determines the loadoriented size of the cuttingfamilies part of are the hardly millingtocutter, However, assembly product find. surface and thelevel, tool life, thus affecting the milling On theroughness product family products differ mainly in two main characteristics: (i) the number components (ii) the processing power, processing qualityofand processingand efficiency, type components (e.g.chip mechanical, electrical, electronical). etc. of The un-deformed parameters have two forms: unClassicalchip methodologies mainly single products deformed thickness andconsidering width of un-deformed chip, which or already existing families analyze the aresolitary, mainly influenced by the product feed, milling depth and cutter product structure on a physical level (components level) which causes difficulties regarding an efficient definition and comparison of different product families. Addressing this
Peer-review under responsibility of the scientific committee of the 51st CIRP Conference on Manufacturing Systems.
2212-8271©©2017 2018The The Authors. Published by Elsevier 2212-8271 Authors. Published by Elsevier B.V. B.V. Peer-review under responsibility of scientific the scientific committee theCIRP 51stDesign CIRP Conference Conference2018. on Manufacturing Systems. Peer-review under responsibility of the committee of the of 28th 10.1016/j.procir.2018.03.075
2
Yongbing Wang et al. / Procedia CIRP 72 (2018) 1380–1385 Yongbing Wang/ Procedia CIRP 00 (2018) 000–000
geometry of the cutting parameters [10].The rational choice of un-deformed chip parameters has significant influence on improving machining quality, production efficiency and reduction of milling energy consumption, so it is necessary to study the effect of un-deformed chip parameters on milling energy consumption [14]. However, the research on the influence of un-deformed chip parameters on the energy consumption of milling is relatively rare. J Gawlik presented a new method to describing chip formation in machining process, which can confirm the minimum thickness of the cutting layer based on acoustic emission signals generated from the cutting area [12].And Li proposed a model of instantaneous undeformed chip thickness, which include the dynamic modulation caused by the tool vibration, and took the dynamic regeneration effect into account. The result shows that tool tip vibration, the cutting force and stability limits of a milling process are predicated [15]. And Habibi Mohsen presented a Semi-analytical representation of the projection, which is based on the un-deformed chip on the rake face of the cutting blades. Then, by using the derived chip geometry and converting facehobbling into oblique cutting, the chip geometry is derived and cutting forces are predicted during face-hobbling by implementing oblique cutting theory [16]. In order to investigate the relationship between the tool run-out and the cutting force, Zhu presented an improved instantaneous milling force per tooth which includes of tool run-out effect. Then the un-deformed chip thickness considering tool run-out are defined and modelled. The results shows that the model can accurately describe the instantaneous force [17]. Most of the literature above take the un-deformed chip parameters calculation method, the related model building and the influence law of the un-deformed chip parameters on cutting force analysis as the research object. It is rarely involves the study of un-deformed chip parameters on the energy consumption of milling. Thus, in this paper take the mutual influence relationship between un-deformed chip parameters and the milling energy consumption as the research object. 2. Problem description and method 2.1 Scope and definition The purpose of this work is to build up a model to show the relationship between the un-deformed chip parameters and the energy consumption in milling process. And the proposed energy consumption model can provides a theoretical basis for mass production of the workpiece in machining systems to reduce the energy consumption. In the process of end milling, due to the milling tool has multiple cutting edges, the milling process can be considered as the cutting process after the previous cutter tooth processing, and the cutting layer between the adjacent two blades constitute the milling cutting layer. And the un-deformed chip parameters include un-deformed chip thickness and width of uncut chip. Cutting parameters and the geometry angle of the cutter will greatly affect un-deformed chip thickness and width of uncut chip. A schematic representation of the un-deformed chip thickness and width of uncut chip is given in Fig.1.
1381
Fig.1. Cutting thickness of face milling cutter
2.2 Effect of un-deformed chip parameters Among various kinds of machining, the cutting metal layer is known as the cutting layer, which is between adjacent two transition surfaces with a layer of metal. The shape and size of the cutting layer directly determine the size of the load and the shape and size of the cutter. In the CNC machining process, undeformed chip parameters determines the load size of the cutting workpiece of the milling cutter, surface roughness and the tool life. Therefore, un-deformed chip significantly affect the milling processing power, processing quality and processing efficiency. 2.2.1 Calculation of un-deformed chip thickness As one of the chip parameters, chip thickness is the distance between the two adjacent blades of the cutting edge. The cutting thickness of a knife teeth is minimal when the workpiece is first cut, and the largest in the middle. Then gradually decreases. Expression of the chip thickness at point i is ℎ𝐷𝐷𝐷𝐷 = 𝑓𝑓𝑍𝑍 𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝐾𝐾𝑟𝑟 .Where 𝒇𝒇𝒁𝒁 is the feed rate in mm/tooth of the milling cutter, θ is the cutting edge angle of the milling cutter, and 𝐾𝐾𝑟𝑟 is tool cutting edge angle[18]. For milling process, the feed rate per tooth ( 𝒇𝒇𝒁𝒁 ) is a constant during the machining. In order to calculate the chip thickness under a constant per tooth feed rate, the average thickness of the non-deformation cutting layer is used as the un-deformed chip thickness under the constant feed rate per tooth. Boothroyd indicated that the average thickness of the non-deformation cutting layer was calculated as ℎ𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷 = 𝑓𝑓𝑍𝑍 𝜑𝜑𝑠𝑠 ∫ 𝑠𝑠𝑠𝑠𝑠𝑠∅𝑑𝑑∅ [19]. 𝜑𝜑𝑆𝑆 0
Where ℎDavg is the average thickness of
the un-deformation cutting layer, and measured in mm. 𝜑𝜑𝑠𝑠 is the tool cutting edge angle in radian, ∅ is the tool cutting edge angle in degree. Therefore, the un-deformed chip thickness in the milling process can be obtained as shown in Eq. (1). ℎ𝐷𝐷 = ℎ𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷 =
𝑓𝑓𝑍𝑍 𝜑𝜑𝑠𝑠 ∫ 𝑠𝑠𝑠𝑠𝑠𝑠∅𝑑𝑑∅ 𝜑𝜑𝑆𝑆 0
(1)
2.2.2 Calculation of the width of the un-deformed chip The width of the un-deformed chip is the contact length between cutting edge and workpiece cutting surface, it is the length of major cutting edge actually participates in the cutting. And 𝑏𝑏𝐷𝐷 been used to present the width of uncut chip, and shown as Eq. (2) 𝑎𝑎 𝑏𝑏𝐷𝐷 = 𝑝𝑝 (2) sin 𝐾𝐾𝑟𝑟
The milling cutter used in this paper belong to the cylindrical straight teeth milling cutter, so the width of undeformed chip as same as the milling width [18]. As shown in Eq. (3). 𝑏𝑏𝐷𝐷 = 𝑎𝑎𝑝𝑝 (3) Where 𝑎𝑎𝑝𝑝 is the cutting depth of milling.
1382
Yongbing Wang et al. / Procedia CIRP 72 (2018) 1380–1385 Yongbing Wang / Procedia CIRP 00 (2018) 000–000
2.3 Energy modelling based on RSM In this paper, the relationship between un-deformed chip parameters and milling energy consumption is analysed by the response surface method (RSM). RSM consists of statistical techniques and a mathematical group and used in the development of an adequate functional relationship between input parameters (cutting conditions) and the output variables (energy consumption) [20]. By establishing the response surface of the regression equation, the fastest ascending path can be obtained, so that the response can reach the optimal value [21]. Compared with other traditional methods, the establishment of the RSM can effectively build-up the complex multidimensional space and closer to the actual situation. By using the method of nonlinear regression fitting response surface model, the model accuracy and study the interaction between the input variables can be effectively improved. In order to study the impact of un-deformed chip parameters on milling energy consumption, RSM usually adopts quadratic regression equation [22]. As shown in Eq. (4). 𝒎𝒎 𝒎𝒎 𝟐𝟐 𝒀𝒀 = 𝜷𝜷𝟎𝟎 + ∑𝒎𝒎 (4) 𝒊𝒊=𝟏𝟏 𝜷𝜷𝒊𝒊 𝒙𝒙𝒊𝒊 + ∑𝒊𝒊<𝒋𝒋 𝜷𝜷𝒊𝒊𝒊𝒊 𝒙𝒙𝒊𝒊 𝒙𝒙𝒋𝒋 + ∑𝒊𝒊=𝟏𝟏 𝜷𝜷𝒊𝒊𝒊𝒊 𝒙𝒙𝒊𝒊 + 𝜺𝜺 Where Y is the response variable, 𝑋𝑋𝑖𝑖 (i=1, 2,…,m) for the cutting layer parameters, 𝛽𝛽0 , 𝛽𝛽𝑖𝑖 , 𝛽𝛽𝑖𝑖𝑖𝑖 , 𝛽𝛽𝑖𝑖𝑖𝑖 are the quadratic regression coefficients, and ε is the error of regression value and actual value. 3. Object of analysis and results This case not only illustrates the process of energy model establishment, but also demonstrates and analysis the influence of un-deformed chip parameters on machine tool energy consumption. Hence, the case analysis includes two aspects: (i) object of analysis; (ii) experiment results. 3.1 Object of analysis This case is to establish the energy consumption model by RSM, and the date is got by the test. The method was to undertake cutting tests at set values of un-deformed chip parameters and evaluate the coefficient of Eq.5. From the relevant experimental requirements in the process of NC milling and according to the processing requirements, the width is 5mm can be chosen in this test. And it can be obtained that when milling the aluminium alloy with high speed steel, the feed f𝑍𝑍 of each tooth is 0.065~0.078mm/z, and the value of the un-deformed chip thickness ℎ𝐷𝐷 in the milling process can be calculated by ℎ𝐷𝐷𝐷𝐷 = 𝑓𝑓𝑍𝑍 𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝐾𝐾𝑟𝑟 , which is 0.041~0.050mm. In the case of aluminium alloy milling, the value range of back cutting depth 𝑎𝑎𝑝𝑝 is 2~ 4mm, and the value range of the width of un-deformed chip 𝑏𝑏𝐷𝐷 is 2~ 4mm[18]. Therefore, the value range of un-deformed chip thickness ℎ𝐷𝐷 in the milling process is 0.041~0.050 mm, the value range of the cutting layer width 𝑏𝑏𝐷𝐷 is 2~ 4mm, and the value range of the feeding speed 𝑣𝑣𝑓𝑓 is 250~ 400mm/min. From the above analysis, this article focuses on the influence of cutting parameters (ℎ𝐷𝐷 ,𝑏𝑏𝐷𝐷 ) on CNC milling the energy consumption. And Central Composite Design (CCD) and response surface method is used to approach on experimental design in this paper. The design parameters of the CCD test are as follows [23]: P= 2, r= 1.414, n= 13, 𝑚𝑚0 = 5, 𝑚𝑚𝑐𝑐 = 𝑚𝑚𝑟𝑟 = 4, the cutting layer thickness ℎ𝐷𝐷 is 𝑥𝑥1 , the cutting layer width 𝑏𝑏𝐷𝐷 is 𝑥𝑥2 , and the cutting layer parameter is encoded
3
by the second-order response surface design level coding formula. The coding formula is shown in Eq. (5). (𝑍𝑍1 +𝑍𝑍2 )
𝑥𝑥𝑗𝑗 =
𝑍𝑍2𝑗𝑗 −𝑍𝑍0𝑗𝑗
(𝑍𝑍2 −𝑍𝑍1 )
∆𝑗𝑗
(5)
Where𝑍𝑍𝑗𝑗0 = ; ∆j = . 2 2 The un-deformed chip parameters are encoded according to formula (5), and the corresponding un-deformed chip parameters in the coding space are shown in table 1. Table 1. Binary response surface CCD design level coding table Cutting layer depth Cutting layer width hD x1(mm) bD x2(mm) Upper asterisk arm 1.414 0.052 4.41 Upper level 1 0.050 4 Zero level 0 0.046 3 Lower level -1 0.041 2 Lower asterisk arm -1.414 0.039 1.59
3.2 Results As an important part of industrial production, mechanical manufacture has a significant impact on energy utilization. Considering that the un-deformed chip parameters have a great influence on the energy consumption of machine tool, the undeformed chip thickness and the width of the un-deformed chip are chosen as the research object in this paper. After milling on the JTVC -650b CNC machining centre, the power demand was calculated from the measured current and the relevant parameters of each experiment [24]. The proposed energy consumption model, which is a tool for evaluating the energy consumption ， and which is constructive with guidance function to carry out the further research [25]. In this study, Central Composite Design (CCD) and response surface method were used to approach on experimental design and energy consumption model. The application of this model can provide some constructive guidance for the machining process of machine tools, and the relationship between parameters and energy consumption of the machine tool based on variance analysis is given.，which can provide a reference for volume production. The method flow chart is shown in Fig. 2. Cutting layer parameters
Experiment setup
Experimental equipment Workpiece materials and the processing requirements Milling cutter CNC milling energy consumption acquisition device
Experimental procedure
Experiment and get the data according to the experiment.
prediction model
Based on the above experimental data，utilize nonlinear regression analysis get model.
Variance analysis of milling energy consumption model Dimension reduction analysis based on milling energy consumption
Discussions and analysis
Fig.2. The method flow chart of the model
3.2.1 Experiment setup Experimental equipment: this experiment takes the milling plane as an example. The machining process is shown in Fig. 3. The machine model is JTVC -650b CNC machining centre, and the value range of spindle speed range is 80~5000 r/min. Meanwhile the value range of feed speed is 80~15000 mm/min, and the more information about the machine tool can be listed as follows: the maximum feed rate 𝑓𝑓𝑚𝑚𝑚𝑚𝑚𝑚 = 1000𝑁𝑁 , the maximum torque 𝑀𝑀𝑚𝑚𝑚𝑚𝑚𝑚 = 200𝑁𝑁 ∙ 𝑚𝑚 , the maximum power 𝑃𝑃𝑚𝑚𝑚𝑚𝑚𝑚 = 7.5 𝑘𝑘𝑘𝑘, the efficiency=0.8.
4
Yongbing Wang et al. / Procedia CIRP 72 (2018) 1380–1385 Yongbing Wang/ Procedia CIRP 00 (2018) 000–000
Manual change of knife button
Spindle
Cooling pipe
End mill
workbench
workpiece Fixture
Fig.3. Experimental setup for the tests
(2) Workpiece materials and the processing requirements: workpiece material is 6061 aluminium alloy, the basic size (length×width×height) for 120×40×50 mm, its mechanical properties are listed as follows: yield strength≥95Mpa, Strength of extension≥240Mpa, elongation≥12%, Vickers hardness≤200. Rough milling shape processing surface quality after Ra is less than 6.3 microns, the machining trials were conducted under a dry condition in order to avoid the influence of cutting fluid. (3) Milling cutter: cylindrical high speed steel end-milling cutter, cutter related parameters are shown in table 2. Table 2. Parameters of milling cutters D (mm) 16
Z 3
β (°) 45
γ (°) 5
∅ (°) 90
𝐿𝐿𝑐𝑐 (mm) 30
L (mm) 75
(4) CNC milling energy consumption acquisition device This experiment adopts PW6001 (Hioki high precision power analyser) to monitor the related milling energy consumption, and uses 3P3W3A wiring method to connect it to the power input of CNC machine tool. The monitoring device is mainly used to measure the consumption of electric energy in the machining process of CNC machine tool with different parameters of the cutting layer, and its connection to the CNC machine is shown in the Fig. 4. Before the test, according to the performance of CNC machine tools, the basic measurement parameters of power analyser are set. It include: machine tool connection mode, voltage range, current range, record time interval, etc., and relevant parameters are set as shown in table 3.
Fig.4. Energy consumption acquisition platform of NC milling
3.2.2 Experimental procedure According to table 1, the test design is obtained in table 4. The table include the test number, the value of the test design
1383
factor in the coding space and the value of the cutting parameters (n,𝑓𝑓𝑍𝑍 ,𝑎𝑎𝑝𝑝 and𝑎𝑎𝑒𝑒 ). In order to study the influence of un-deformed chip parameters and milling energy consumption, here we select spindle speed and milling width as fixed values. And through the width of un-deformed chip and tooth feed got different cutting speed, then take cutting thickness and width of uncut chip of cutting speed and each tooth feed conversion as milling processing variables. The energy consumption of milling under different cutting parameters value is obtained by experiments as shown in table 4. 3.2.3 Construction of milling energy consumption prediction model Based on the above experimental data, the data processing and nonlinear regression analysis of the experimental results in table 4 are carried out by using the data statistical analysis software design-expert 8.0. The response surface model of the milling energy consumption and un-deformed chip parameters in the coding state is obtained by the nonlinear quadratic regression analysis, as shown in Eq. (6). 𝐸𝐸 = 2.03 − 0.47𝑥𝑥1 + 0.067𝑥𝑥2 + 0.27𝑥𝑥1 𝑥𝑥2 + 0.022𝑥𝑥12 − 0.12𝑥𝑥22 (6) In the model, the model is under the fixed spindle speed n and cutting width 𝑎𝑎𝑒𝑒 , thus the analysis and conclusions of the model are set up in the same spindle speed and the processing conditions of cutting width. And this model is used to study influence on the cutting power under different parameters (ℎ𝐷𝐷 , 𝑏𝑏𝐷𝐷 ) of un-deformed chip. Order to get the influence law of two parameters (ℎ𝐷𝐷 , 𝑏𝑏𝐷𝐷 ) on actual milling energy consumption, the response surface model of the coding state is decoded according to the coding formula (5). And then the actual quadratic function model of un-deformed chip parameters and milling energy consumption is obtained. As is shown in Eq. (7) 𝐸𝐸 = 8.879 − 221.834ℎ𝐷𝐷 + 0.474𝑏𝑏𝐷𝐷 + 6.11ℎ𝐷𝐷 × 𝑏𝑏𝐷𝐷 + 1098.765ℎ𝐷𝐷2 − 0.115𝑏𝑏𝐷𝐷2 (7) 4. Discussions 4.1 Variance analysis of milling energy consumption model In order to analyse the relative influence between undeformed chip parameters and milling energy consumption, and to verify the accuracy of milling energy fitting model, Analysis of Variance( ANOVA) is used to analyse the milling energy consumption [26]. The variance analysis of the milling energy consumption response surface model is shown in table 5. In general, F>4 indicates that the change of design variable has a significant effect on response variables. From table 5, the F ratio of the milling energy consumption model is 75.33, it’s far larger than 4. So it can be concluded that the proposed quadratic regression model have a highly significant and the prediction accuracy is high. Meanwhile, this paper selects α =0.05 as the confidence level to determine the statistical significance of the results. When p< 0.05, it indicates that the significance of the corresponding item in the fitting model is extremely high. As is shown in table 5, the p of the fitting model is less than 0.0001, which means the statistical significance is very high, and the significance of the various dimensions in the model areℎ𝐷𝐷 , 𝑏𝑏𝐷𝐷 2 , 𝑏𝑏𝐷𝐷 , ℎ𝐷𝐷 2 and ℎ𝐷𝐷 × 𝑏𝑏𝐷𝐷 , respectively[27]. So in the milling process, un-deformed chip thickness is the main factor affecting the energy consumption
Yongbing Wang et al. / Procedia CIRP 72 (2018) 1380–1385 Yongbing Wang / Procedia CIRP 00 (2018) 000–000
1384
5
Table 3. PW6001 power analyser parameter setting Setting up the project
Connection mode
Voltage range
Current range
Current sensor mode
voltage conversion ratio （VT） 1.0
set value 3P3W3A 1~600V 1~60A 9707 Table 4. Design and result of the combined test design of surface centre Experimental design Implement of experiment Test number n/r.min-1 vf/mm/min ℎ𝐷𝐷 /mm 𝑏𝑏𝐷𝐷 /mm x1 𝑥𝑥2 1 -1.00 -1.00 0.041 2.00 2000 250 2 1.00 -1.00 0.050 2.00 2000 400 3 -1.00 1.00 0.041 4.00 2000 250 4 1.00 1.00 0.050 4.00 2000 400 5 -1.41 0.00 0.039 3.00 2000 219 6 1.41 0.00 0.052 3.00 2000 431 7 0.00 -1.41 0.046 1.59 2000 325 8 0.00 1.41 0.046 4.41 2000 325 9 0.00 0.00 0.046 3.00 2000 325 10 0.00 0.00 0.046 3.00 2000 325 11 0.00 0.00 0.046 3.00 2000 325 12 0.00 0.00 0.046 3.00 2000 325 13 0.00 0.00 0.046 3.00 2000 325 Table 5. Analysis of the variance of the energy response surface model of the milling Sum of squares of degree of Source variance of mean variance freedom Regression model 1.87 5 0.37 hD 1.74 1 1.74 bD 0.030 1 0.030 hD×bD 0.00303 1 0.00303 hD² 0.00344 1 0.00344 bD² 0.092 1 0.092 Residual Error 0.035 7 0.00497 Lack of fit 0.022 3 0.00750 Pure Errol 0.012 4 0.00307 Total 1.91 12
of milling, followed by width of uncut chip. When p>0.05, it indicates that the corresponding item in the fitting model is not significant. Therefore, the results of the analysis and table 5 indicate that the proposed quadratic regression model have a higher significance and its model can predict the milling energy consumption value of different cutting layer parameters. Meanwhile, the values of R – sq is one of the important indexes of fitting the model. It can indicate that the model is good or bad and makes a better performance of the model predicts. And when its value is close to 1, shows that the fitting quadratic response surface model prediction of the closer the value and the actual data. From table 5, the value of R - sq (0.9818) and R - sq (adjust) (0.9687) are close to the index 1. It shows that the fitting degree of milling energy consumption model is good, and predicted range of values of energy consumption is effectively. According to the experimental data and the fitting model, the normal residual distribution is shown in Fig. 5. From Fig. 5, the coincidence between predicted and measured values is high, almost straight, and the error is small. Taking the undeformed chip parameters of group 5 and group 7 as an example, the E5 and E7 was obtained as E5=3.02, E7=1.93. Compared with the measured value, Est5= 3.09, Est7= 1.99, the error is 2.27% and 3.02% respectively. And the error is within 5%, it is indicated that the experimental model and measured value fit well. Thus the model based on nonlinear quadratic regression fits the milling energy consumption prediction is very well. And the influence of input variables and response variables can be effectively mirrored by the model.
Current conversion ratio （CT） 1.0
ap/mm 2.00 2.00 4.00 4.00 3.00 3.00 1.59 4.41 3.00 3.00 3.00 3.00 3.00
ae/mm 4 4 4 4 4 4 4 4 4 4 4 4 4
Data record interval 500ms Experimental result Est/wh 2.58 1.68 2.62 1.83 3.09 1.65 1.99 2.20 2.33 2.32 2.20 2.33 2.31
F ratio
P
75.33 349.50 5.97 0.61 0.69 18.60
< 0.0001 < 0.0001 0.0446 0.4607 0.4325 0.0035
significant
2.44
0.2042
not significant
Fig.5. (a) The normal probability graph of residuals; (b) The relation diagram of the predicted value and the actual value.
4.2 Dimension reduction analysis based on milling energy consumption According to the quadratic regression equation, the corresponding surface map can be drawn. The variable factor of this experiment is the un-deformed chip thickness and the dimension reduction plane between the cutting parameters and the energy consumption of milling is obtained through the reduction of dimension. From them, the relationship between the un-deformed chip parameters ( ℎ𝐷𝐷 , 𝑏𝑏𝐷𝐷 ) and the energy response variables of the milling energy consumption is obtained. As is shown in Fig. 6.
Fig.6.Un-deformed chip thickness and width of uncut chip response surface
6
Yongbing Wang et al. / Procedia CIRP 72 (2018) 1380–1385 Yongbing Wang/ Procedia CIRP 00 (2018) 000–000
From the figure above, it can be drawn that when the undeformed chip thickness and width of uncut chip are large, the energy consumption of milling is small. When in the actual milling process, if the un-deformed chip thickness and width of uncut chip are larger, the milling time will become shorter. And meanwhile under the CNC machining conditions of stability and power under the condition of small fluctuations, it will make the machine total energy consumption small. Take the group 4, 6, 8 as example, the experimental results show that the larger un-deformed chip thickness and the larger width of uncut chip can reduce the energy consumption of the milling process. 5．Conclusion In this paper, the test design of the un-deformed chip parameters was carried out by the response surface centre combination method (CCD). And the regression model of the un-deformed chip parameters ( ℎ𝐷𝐷 , 𝑏𝑏𝐷𝐷 ) and milling energy consumption was established by using nonlinear quadratic regression equation. Based on the results above, the following conclusions can be drawn: (1)The analysis of the fitting model shows that the model can be used to analyse and predict the influence of un-deformed chip parameters on the energy consumption of milling. And the milling energy consumption decreases with the increase of the un-deformed chip thickness and the width of uncut chip. (2)When in the actual milling process, if the un-deformed chip thickness and width of uncut chip are larger, the milling time will become shorter. In order to obtain the lower milling energy consumption, it is important to select the larger undeformed chip parameters and shorten the cutting time. References [1]
International Energy Agency (IEA). Tracking industrial energy efficiency and CO2 emission [homepage on the Internet]. c2007 [updated 2007 June; cited 09.10.13]. http://www.iea.org/publications/freepublications/publication/tracking_ emissions.pdf [2] Andrea Trianni, Enrico Cagno, Stefano Farné. Barriers, drivers and decision-making process for industrial energy efficiency: A broad study among manufacturing small and medium-sized enterprises. Applied Energy.2016; 162:1537–1551. [3] International Energy Agency (IEA). International energy outlook2017. 2017 https://www.eia.gov/pressroom/presentations/mead_91417.pdf [4] Made in china 2025, 2015. http://www.agri.cn/V20/SC/jjps/201505/t20150520_4605792.htm [5] Li L, Deng XG., et al. Multi-objective optimization of tool path considering efficiency, energy-saving and carbon-emission for freeform surface milling. Journal of Cleaner Production 2017;172:33113322 [6] Liu F, Zhou XN, Cai W. The complexity and strategy for establishing product energy consumption allowance in discrete manufacturing industry. J Mech Eng 2015; 10 (19):138-45 [In Chinese]. [7] Gutowski T G, Branham M S, Dahmus J B. Thermodynamic Analysis of Resources Used in Manufacturing Processes. Environmental Science and Technology 2009; 43(5):1584–1590. [8] Pusavec F, Kramar D, Krajnik P et al. Transitioning to sustainable production – part II: evaluation of sustainable machining technologies. Journal of Cleaner Production, 2010, 18(12):1211-1221. [9] Zhao G.Y., et al. Energy Consumption in Machining: Classification, Prediction, and Reduction Strategy. Energy 2017; 133. [10] Balogun V A., P. T. Mativenga. Impact of un-deformed chip thickness on specific energy in mechanical machining processes. Journal of Cleaner Production 2014; 69.69:260-268. [11] Jia S., et al. Therblig-embedded value stream mapping method for lean energy machining. Journal of Cleaner Production;2017; 138:1081-1098 [12] J. Gawlik et al., Identification of the Minimal Thickness of Cutting
[13]
[14] [15] [16] [17]
[18] [19] [20]
[21] [22] [23] [24] [25] [26] [27]
1385
Layer Based on the Acoustic Emission Signal. Key Engineering Materials 2016; 686:39-44. Krajnik P, et al. transitioning to sustainable production – part III: developments and possibilities for integration of nanotechnology into material processing technologies. Journal of Cleaner Production 2015; 2010.1:204-204. Cai W, Liu F, Zhou X N. Fine energy consumption allowance of workpieces in the mechanical manufacturing industry. Energy 2016; 114: 622–3. Li Z Q, S. Zhang. Cutting force modeling and simulation of tee slot milling. International Conference on Mechatronics and Automation IEEE, 2009:1438-1443. Habibi Mohsen, Z. C. Chen. A semi-analytical approach to un-deformed chip boundary theory and cutting force prediction in face-hobbling of bevel gears. Computer-Aided Design 2016; 73:53-65. Zhu Kunpeng, Y. Zhang. Modeling of the instantaneous milling force per tooth with tool run-out effect in high speed ball-end milling. International Journal of Machine Tools & Manufactures 2017; 118– 119:37-48. Chen XP, He XN. Metal-cutting principles and tools. Peking University Press; Beijing; 2006 [in Chinese] Boothroyd, G., W. A. Knight. Fundamentals of Machining and Machine Tools. Marcel Dekker, 1989. Nouioua M, Yallese M A, Khettabi R, et al. Investigation of the performance of the MQL, dry, and wet turning by response surface methodology (RSM) and artificial neural network (ANN). International Journal of Advanced Manufacturing Technology 2017; (2):1-20. Fnides B, Yallese M A, Mabrouki T, et al. Application of response surface methodology for determining cutting force model in turning hardened AISI H11 hot work tool steel. Sādhanā, 2011, 36(1):109-123. Thepsonthi T, Özel T. Multi-objective process optimization for microend milling of Ti-6Al-4V titanium alloy. The International Journal of Advanced Manufacturing Technology, 2012, 63(9):903-914. Cook Dennis, Fedorov Valery. Constrained Optimization of Experimental Design. Statistics 1995; 26.2:129--178. Schudeleit Timo, S. Züst, K. Wegener. Methods for evaluation of energy efficiency of machine tools. Energy 2015; 93:1964-1970. Cai Wei, et al. A tool for assessing the energy demand and efficiency of machining systems: Energy benchmarking. Energy 2017; 138:332-347. Anderson, Marti J. A new method for non-parametric multivariate analysis of variance. Austral Ecology 2010; 26.1:32-46. Hu Luoke, et al. Minimising the machining energy consumption of a machine tool by sequencing the features of a part. Energy 2017; 121:292–305.
ScienceDirect
Available online atonline www.sciencedirect.com Available at www.sciencedirect.com
ScienceDirect ScienceDirect
Procedia CIRP 00 (2018) 000–000
www.elsevier.com/locate/procedia
ProcediaProcedia CIRP 00CIRP (2017) 72000–000 (2018) 1380–1385 www.elsevier.com/locate/procedia
51st CIRP Conference on Manufacturing Systems
Exploring the of un-deformed parameters 28th effect CIRP Design Conference, Maychip 2018, Nantes, France on energy consumption for energy efficiency improvement in the milling A new methodology toYongbing analyze the physical architecture of Wang , Lifunctional li *, Li Lingling and Wei Cai , College of engineering and technology of southwest university, Chongqing,400715, China existing products for an assembly oriented product family identification State Key Laboratory of Mechanical Transmission, Chongqing University, Chongqing, 400030, China a
a
a
b
a
b
Paul Stief *, Jean-Yves Dantan, Alain Etienne, Ali Siadat
* Corresponding author. Tel.: +86-150-2309-2936; fax: +86 023 98251265. E-mail address: [email protected]
École Nationale Supérieure d’Arts et Métiers, Arts et Métiers ParisTech, LCFC EA 4495, 4 Rue Augustin Fresnel, Metz 57078, France
* Corresponding author. Tel.: +33 3 87 37 54 30; E-mail address: [email protected]
Abstract In order to study the influence of milling energy consumption and un-deformed chip parameters, the energy consumption model is Abstract constructed and the validity of the model is analyzed. Firstly, the relationship between un-deformed chip parameters and cutting
is analyzed. Then RSM andtowards centralmore composite is used to get the model. Bytousing the multiple the regression Inparameters today’s business environment, the trend productdesign varietymethod and customization is unbroken. Due this development, need of agile and reconfigurable production to copesurface with various and and product families. To design and optimize production analysis on the experimental datasystems fitting,emerged the response modelproducts of milling cutting parameters in energy consumption is systems as well as toby choose optimalanalysis, product matches, product are needed. of the known methods to obtained. Finally, usingthe variance the model showsanalysis that themethods experimental fittingIndeed, modelmost significantly higher, at theaim same analyze a product one productreduction family on of theun-deformed physical level.chip Different producton families, however, may differ of largely in terms the number and time through theordimension parameters the energy consumption milling planeofwere analyzed. nature of components. This fact impedes an efficient comparison and choice of appropriate product family combinations for the production system. newAuthors. methodology is proposed to analyze © 2018AThe Published by Elsevier B.V. existing products in view of their functional and physical architecture. The aim is to cluster these productsunder in new assembly oriented product families for the optimization existing assembly lines and the creation of future reconfigurable Peer-review responsibility of the scientific committee of the 51st CIRPofConference on Manufacturing Systems. assembly systems. Based on Datum Flow Chain, the physical structure of the products is analyzed. Functional subassemblies are identified, and a Keywords: functional analysis performed. Moreover, a hybrid functionalchip and parameters; physical architecture graph (HyFPAG) is Analysis the outputofwhich depicts the Energyisconsumption of milling; Un-deformed Central composite design; variance; similarity between product families by providing design support to both, production system planners and product designers. An illustrative Dimension reduction example of a nail-clipper is used to explain the proposed methodology. An industrial case study on two product families of steering columns of thyssenkrupp Presta France is then carried out to give a first industrial evaluation of the proposed approach.potential [6][7]. In the process of considerable energy-saving 1. Introduction © 2017 The Authors. Published by Elsevier B.V. manufacture of machine tool structure design, product design Peer-review under responsibility of the scientific committee of the 28th CIRP Conference 2018.optimization, it can effectively reduce andDesign machine parameters
CNC processing, as the main production mode of manufacturing, widely used in identification the machine process of Keywords: Assembly;isDesign method; Family industrial products. However, as the rise in energy prices and awareness of environmental protection, the reduction of electricity consumption and CO2 emission become the driving 1.force Introduction for optimizing energy demand in manufacturing industry [1][2]. The U.S. Energy Information Administration (EIA) Due toan the fast development in2017, the and domain of published International energy outlook the global communication and an ongoing trend of digitization and energy-related carbon dioxide (CO2)emissions are expected digitalization, enterprises facing important billion to rise from manufacturing 33 billion metric tons inare 2015 to 39 challenges in today’s market environments: a continuing metric tons in 2040[3]. Meanwhile, the china government has tendency towards reduction oftoproduct development and also made a long-term policy reducing the carbon times emissions shortened product lifecycles. In addition, there is an increasing [4]. The relative standard they have set is to reduce emissions demand of customization, at the same a global of a particulate matter being by 40%by 2040.time It in forces the competition with competitors all over the world. This manufacturing industry to face the development oftrend, high which is inducing the and development from macro to [5][3]. micro efficiency, high speed low energy consumption markets, results in diminished lot sizes due to augmenting With a wide distribution and large amount of energy product varieties to low-volume production) [1]. consumption at (high-volume a low efficiency, machining process have To cope with this augmenting variety as well as to be able to identify possible optimization potentials in the existing 2212-8271 ©system, 2018 TheitAuthors. Published Elsevier B.V. knowledge production is important tobyhave a precise
the energy consumption, product manufacturing process is the effective way to realize the green of manufacturing high efficiency, low energy consumption and sustainable development [8][13]. For better analysis and evaluate the energy consumption of characteristics the production process ofand/or NC of the product range and manufactured machininginsystem, it is necessary to model and challenge analyse the assembled this system. In this context, the main in energy consumption of CNC machining under different modelling and analysis is now not onlysystem to cope with single technological parameters and or establish the relationship products, a limited product range existing product families, between parameters and products process to energy but also to beprocess able to analyze and to compare define consumption[9][11]. new product families. It can be observed that classical existing In the process NC machining, un-deformed directly product families areofregrouped in function of clientschip or features. determines the loadoriented size of the cuttingfamilies part of are the hardly millingtocutter, However, assembly product find. surface and thelevel, tool life, thus affecting the milling On theroughness product family products differ mainly in two main characteristics: (i) the number components (ii) the processing power, processing qualityofand processingand efficiency, type components (e.g.chip mechanical, electrical, electronical). etc. of The un-deformed parameters have two forms: unClassicalchip methodologies mainly single products deformed thickness andconsidering width of un-deformed chip, which or already existing families analyze the aresolitary, mainly influenced by the product feed, milling depth and cutter product structure on a physical level (components level) which causes difficulties regarding an efficient definition and comparison of different product families. Addressing this
Peer-review under responsibility of the scientific committee of the 51st CIRP Conference on Manufacturing Systems.
2212-8271©©2017 2018The The Authors. Published by Elsevier 2212-8271 Authors. Published by Elsevier B.V. B.V. Peer-review under responsibility of scientific the scientific committee theCIRP 51stDesign CIRP Conference Conference2018. on Manufacturing Systems. Peer-review under responsibility of the committee of the of 28th 10.1016/j.procir.2018.03.075
2
Yongbing Wang et al. / Procedia CIRP 72 (2018) 1380–1385 Yongbing Wang/ Procedia CIRP 00 (2018) 000–000
geometry of the cutting parameters [10].The rational choice of un-deformed chip parameters has significant influence on improving machining quality, production efficiency and reduction of milling energy consumption, so it is necessary to study the effect of un-deformed chip parameters on milling energy consumption [14]. However, the research on the influence of un-deformed chip parameters on the energy consumption of milling is relatively rare. J Gawlik presented a new method to describing chip formation in machining process, which can confirm the minimum thickness of the cutting layer based on acoustic emission signals generated from the cutting area [12].And Li proposed a model of instantaneous undeformed chip thickness, which include the dynamic modulation caused by the tool vibration, and took the dynamic regeneration effect into account. The result shows that tool tip vibration, the cutting force and stability limits of a milling process are predicated [15]. And Habibi Mohsen presented a Semi-analytical representation of the projection, which is based on the un-deformed chip on the rake face of the cutting blades. Then, by using the derived chip geometry and converting facehobbling into oblique cutting, the chip geometry is derived and cutting forces are predicted during face-hobbling by implementing oblique cutting theory [16]. In order to investigate the relationship between the tool run-out and the cutting force, Zhu presented an improved instantaneous milling force per tooth which includes of tool run-out effect. Then the un-deformed chip thickness considering tool run-out are defined and modelled. The results shows that the model can accurately describe the instantaneous force [17]. Most of the literature above take the un-deformed chip parameters calculation method, the related model building and the influence law of the un-deformed chip parameters on cutting force analysis as the research object. It is rarely involves the study of un-deformed chip parameters on the energy consumption of milling. Thus, in this paper take the mutual influence relationship between un-deformed chip parameters and the milling energy consumption as the research object. 2. Problem description and method 2.1 Scope and definition The purpose of this work is to build up a model to show the relationship between the un-deformed chip parameters and the energy consumption in milling process. And the proposed energy consumption model can provides a theoretical basis for mass production of the workpiece in machining systems to reduce the energy consumption. In the process of end milling, due to the milling tool has multiple cutting edges, the milling process can be considered as the cutting process after the previous cutter tooth processing, and the cutting layer between the adjacent two blades constitute the milling cutting layer. And the un-deformed chip parameters include un-deformed chip thickness and width of uncut chip. Cutting parameters and the geometry angle of the cutter will greatly affect un-deformed chip thickness and width of uncut chip. A schematic representation of the un-deformed chip thickness and width of uncut chip is given in Fig.1.
1381
Fig.1. Cutting thickness of face milling cutter
2.2 Effect of un-deformed chip parameters Among various kinds of machining, the cutting metal layer is known as the cutting layer, which is between adjacent two transition surfaces with a layer of metal. The shape and size of the cutting layer directly determine the size of the load and the shape and size of the cutter. In the CNC machining process, undeformed chip parameters determines the load size of the cutting workpiece of the milling cutter, surface roughness and the tool life. Therefore, un-deformed chip significantly affect the milling processing power, processing quality and processing efficiency. 2.2.1 Calculation of un-deformed chip thickness As one of the chip parameters, chip thickness is the distance between the two adjacent blades of the cutting edge. The cutting thickness of a knife teeth is minimal when the workpiece is first cut, and the largest in the middle. Then gradually decreases. Expression of the chip thickness at point i is ℎ𝐷𝐷𝐷𝐷 = 𝑓𝑓𝑍𝑍 𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝐾𝐾𝑟𝑟 .Where 𝒇𝒇𝒁𝒁 is the feed rate in mm/tooth of the milling cutter, θ is the cutting edge angle of the milling cutter, and 𝐾𝐾𝑟𝑟 is tool cutting edge angle[18]. For milling process, the feed rate per tooth ( 𝒇𝒇𝒁𝒁 ) is a constant during the machining. In order to calculate the chip thickness under a constant per tooth feed rate, the average thickness of the non-deformation cutting layer is used as the un-deformed chip thickness under the constant feed rate per tooth. Boothroyd indicated that the average thickness of the non-deformation cutting layer was calculated as ℎ𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷 = 𝑓𝑓𝑍𝑍 𝜑𝜑𝑠𝑠 ∫ 𝑠𝑠𝑠𝑠𝑠𝑠∅𝑑𝑑∅ [19]. 𝜑𝜑𝑆𝑆 0
Where ℎDavg is the average thickness of
the un-deformation cutting layer, and measured in mm. 𝜑𝜑𝑠𝑠 is the tool cutting edge angle in radian, ∅ is the tool cutting edge angle in degree. Therefore, the un-deformed chip thickness in the milling process can be obtained as shown in Eq. (1). ℎ𝐷𝐷 = ℎ𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷 =
𝑓𝑓𝑍𝑍 𝜑𝜑𝑠𝑠 ∫ 𝑠𝑠𝑠𝑠𝑠𝑠∅𝑑𝑑∅ 𝜑𝜑𝑆𝑆 0
(1)
2.2.2 Calculation of the width of the un-deformed chip The width of the un-deformed chip is the contact length between cutting edge and workpiece cutting surface, it is the length of major cutting edge actually participates in the cutting. And 𝑏𝑏𝐷𝐷 been used to present the width of uncut chip, and shown as Eq. (2) 𝑎𝑎 𝑏𝑏𝐷𝐷 = 𝑝𝑝 (2) sin 𝐾𝐾𝑟𝑟
The milling cutter used in this paper belong to the cylindrical straight teeth milling cutter, so the width of undeformed chip as same as the milling width [18]. As shown in Eq. (3). 𝑏𝑏𝐷𝐷 = 𝑎𝑎𝑝𝑝 (3) Where 𝑎𝑎𝑝𝑝 is the cutting depth of milling.
1382
Yongbing Wang et al. / Procedia CIRP 72 (2018) 1380–1385 Yongbing Wang / Procedia CIRP 00 (2018) 000–000
2.3 Energy modelling based on RSM In this paper, the relationship between un-deformed chip parameters and milling energy consumption is analysed by the response surface method (RSM). RSM consists of statistical techniques and a mathematical group and used in the development of an adequate functional relationship between input parameters (cutting conditions) and the output variables (energy consumption) [20]. By establishing the response surface of the regression equation, the fastest ascending path can be obtained, so that the response can reach the optimal value [21]. Compared with other traditional methods, the establishment of the RSM can effectively build-up the complex multidimensional space and closer to the actual situation. By using the method of nonlinear regression fitting response surface model, the model accuracy and study the interaction between the input variables can be effectively improved. In order to study the impact of un-deformed chip parameters on milling energy consumption, RSM usually adopts quadratic regression equation [22]. As shown in Eq. (4). 𝒎𝒎 𝒎𝒎 𝟐𝟐 𝒀𝒀 = 𝜷𝜷𝟎𝟎 + ∑𝒎𝒎 (4) 𝒊𝒊=𝟏𝟏 𝜷𝜷𝒊𝒊 𝒙𝒙𝒊𝒊 + ∑𝒊𝒊<𝒋𝒋 𝜷𝜷𝒊𝒊𝒊𝒊 𝒙𝒙𝒊𝒊 𝒙𝒙𝒋𝒋 + ∑𝒊𝒊=𝟏𝟏 𝜷𝜷𝒊𝒊𝒊𝒊 𝒙𝒙𝒊𝒊 + 𝜺𝜺 Where Y is the response variable, 𝑋𝑋𝑖𝑖 (i=1, 2,…,m) for the cutting layer parameters, 𝛽𝛽0 , 𝛽𝛽𝑖𝑖 , 𝛽𝛽𝑖𝑖𝑖𝑖 , 𝛽𝛽𝑖𝑖𝑖𝑖 are the quadratic regression coefficients, and ε is the error of regression value and actual value. 3. Object of analysis and results This case not only illustrates the process of energy model establishment, but also demonstrates and analysis the influence of un-deformed chip parameters on machine tool energy consumption. Hence, the case analysis includes two aspects: (i) object of analysis; (ii) experiment results. 3.1 Object of analysis This case is to establish the energy consumption model by RSM, and the date is got by the test. The method was to undertake cutting tests at set values of un-deformed chip parameters and evaluate the coefficient of Eq.5. From the relevant experimental requirements in the process of NC milling and according to the processing requirements, the width is 5mm can be chosen in this test. And it can be obtained that when milling the aluminium alloy with high speed steel, the feed f𝑍𝑍 of each tooth is 0.065~0.078mm/z, and the value of the un-deformed chip thickness ℎ𝐷𝐷 in the milling process can be calculated by ℎ𝐷𝐷𝐷𝐷 = 𝑓𝑓𝑍𝑍 𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝐾𝐾𝑟𝑟 , which is 0.041~0.050mm. In the case of aluminium alloy milling, the value range of back cutting depth 𝑎𝑎𝑝𝑝 is 2~ 4mm, and the value range of the width of un-deformed chip 𝑏𝑏𝐷𝐷 is 2~ 4mm[18]. Therefore, the value range of un-deformed chip thickness ℎ𝐷𝐷 in the milling process is 0.041~0.050 mm, the value range of the cutting layer width 𝑏𝑏𝐷𝐷 is 2~ 4mm, and the value range of the feeding speed 𝑣𝑣𝑓𝑓 is 250~ 400mm/min. From the above analysis, this article focuses on the influence of cutting parameters (ℎ𝐷𝐷 ,𝑏𝑏𝐷𝐷 ) on CNC milling the energy consumption. And Central Composite Design (CCD) and response surface method is used to approach on experimental design in this paper. The design parameters of the CCD test are as follows [23]: P= 2, r= 1.414, n= 13, 𝑚𝑚0 = 5, 𝑚𝑚𝑐𝑐 = 𝑚𝑚𝑟𝑟 = 4, the cutting layer thickness ℎ𝐷𝐷 is 𝑥𝑥1 , the cutting layer width 𝑏𝑏𝐷𝐷 is 𝑥𝑥2 , and the cutting layer parameter is encoded
3
by the second-order response surface design level coding formula. The coding formula is shown in Eq. (5). (𝑍𝑍1 +𝑍𝑍2 )
𝑥𝑥𝑗𝑗 =
𝑍𝑍2𝑗𝑗 −𝑍𝑍0𝑗𝑗
(𝑍𝑍2 −𝑍𝑍1 )
∆𝑗𝑗
(5)
Where𝑍𝑍𝑗𝑗0 = ; ∆j = . 2 2 The un-deformed chip parameters are encoded according to formula (5), and the corresponding un-deformed chip parameters in the coding space are shown in table 1. Table 1. Binary response surface CCD design level coding table Cutting layer depth Cutting layer width hD x1(mm) bD x2(mm) Upper asterisk arm 1.414 0.052 4.41 Upper level 1 0.050 4 Zero level 0 0.046 3 Lower level -1 0.041 2 Lower asterisk arm -1.414 0.039 1.59
3.2 Results As an important part of industrial production, mechanical manufacture has a significant impact on energy utilization. Considering that the un-deformed chip parameters have a great influence on the energy consumption of machine tool, the undeformed chip thickness and the width of the un-deformed chip are chosen as the research object in this paper. After milling on the JTVC -650b CNC machining centre, the power demand was calculated from the measured current and the relevant parameters of each experiment [24]. The proposed energy consumption model, which is a tool for evaluating the energy consumption ， and which is constructive with guidance function to carry out the further research [25]. In this study, Central Composite Design (CCD) and response surface method were used to approach on experimental design and energy consumption model. The application of this model can provide some constructive guidance for the machining process of machine tools, and the relationship between parameters and energy consumption of the machine tool based on variance analysis is given.，which can provide a reference for volume production. The method flow chart is shown in Fig. 2. Cutting layer parameters
Experiment setup
Experimental equipment Workpiece materials and the processing requirements Milling cutter CNC milling energy consumption acquisition device
Experimental procedure
Experiment and get the data according to the experiment.
prediction model
Based on the above experimental data，utilize nonlinear regression analysis get model.
Variance analysis of milling energy consumption model Dimension reduction analysis based on milling energy consumption
Discussions and analysis
Fig.2. The method flow chart of the model
3.2.1 Experiment setup Experimental equipment: this experiment takes the milling plane as an example. The machining process is shown in Fig. 3. The machine model is JTVC -650b CNC machining centre, and the value range of spindle speed range is 80~5000 r/min. Meanwhile the value range of feed speed is 80~15000 mm/min, and the more information about the machine tool can be listed as follows: the maximum feed rate 𝑓𝑓𝑚𝑚𝑚𝑚𝑚𝑚 = 1000𝑁𝑁 , the maximum torque 𝑀𝑀𝑚𝑚𝑚𝑚𝑚𝑚 = 200𝑁𝑁 ∙ 𝑚𝑚 , the maximum power 𝑃𝑃𝑚𝑚𝑚𝑚𝑚𝑚 = 7.5 𝑘𝑘𝑘𝑘, the efficiency=0.8.
4
Yongbing Wang et al. / Procedia CIRP 72 (2018) 1380–1385 Yongbing Wang/ Procedia CIRP 00 (2018) 000–000
Manual change of knife button
Spindle
Cooling pipe
End mill
workbench
workpiece Fixture
Fig.3. Experimental setup for the tests
(2) Workpiece materials and the processing requirements: workpiece material is 6061 aluminium alloy, the basic size (length×width×height) for 120×40×50 mm, its mechanical properties are listed as follows: yield strength≥95Mpa, Strength of extension≥240Mpa, elongation≥12%, Vickers hardness≤200. Rough milling shape processing surface quality after Ra is less than 6.3 microns, the machining trials were conducted under a dry condition in order to avoid the influence of cutting fluid. (3) Milling cutter: cylindrical high speed steel end-milling cutter, cutter related parameters are shown in table 2. Table 2. Parameters of milling cutters D (mm) 16
Z 3
β (°) 45
γ (°) 5
∅ (°) 90
𝐿𝐿𝑐𝑐 (mm) 30
L (mm) 75
(4) CNC milling energy consumption acquisition device This experiment adopts PW6001 (Hioki high precision power analyser) to monitor the related milling energy consumption, and uses 3P3W3A wiring method to connect it to the power input of CNC machine tool. The monitoring device is mainly used to measure the consumption of electric energy in the machining process of CNC machine tool with different parameters of the cutting layer, and its connection to the CNC machine is shown in the Fig. 4. Before the test, according to the performance of CNC machine tools, the basic measurement parameters of power analyser are set. It include: machine tool connection mode, voltage range, current range, record time interval, etc., and relevant parameters are set as shown in table 3.
Fig.4. Energy consumption acquisition platform of NC milling
3.2.2 Experimental procedure According to table 1, the test design is obtained in table 4. The table include the test number, the value of the test design
1383
factor in the coding space and the value of the cutting parameters (n,𝑓𝑓𝑍𝑍 ,𝑎𝑎𝑝𝑝 and𝑎𝑎𝑒𝑒 ). In order to study the influence of un-deformed chip parameters and milling energy consumption, here we select spindle speed and milling width as fixed values. And through the width of un-deformed chip and tooth feed got different cutting speed, then take cutting thickness and width of uncut chip of cutting speed and each tooth feed conversion as milling processing variables. The energy consumption of milling under different cutting parameters value is obtained by experiments as shown in table 4. 3.2.3 Construction of milling energy consumption prediction model Based on the above experimental data, the data processing and nonlinear regression analysis of the experimental results in table 4 are carried out by using the data statistical analysis software design-expert 8.0. The response surface model of the milling energy consumption and un-deformed chip parameters in the coding state is obtained by the nonlinear quadratic regression analysis, as shown in Eq. (6). 𝐸𝐸 = 2.03 − 0.47𝑥𝑥1 + 0.067𝑥𝑥2 + 0.27𝑥𝑥1 𝑥𝑥2 + 0.022𝑥𝑥12 − 0.12𝑥𝑥22 (6) In the model, the model is under the fixed spindle speed n and cutting width 𝑎𝑎𝑒𝑒 , thus the analysis and conclusions of the model are set up in the same spindle speed and the processing conditions of cutting width. And this model is used to study influence on the cutting power under different parameters (ℎ𝐷𝐷 , 𝑏𝑏𝐷𝐷 ) of un-deformed chip. Order to get the influence law of two parameters (ℎ𝐷𝐷 , 𝑏𝑏𝐷𝐷 ) on actual milling energy consumption, the response surface model of the coding state is decoded according to the coding formula (5). And then the actual quadratic function model of un-deformed chip parameters and milling energy consumption is obtained. As is shown in Eq. (7) 𝐸𝐸 = 8.879 − 221.834ℎ𝐷𝐷 + 0.474𝑏𝑏𝐷𝐷 + 6.11ℎ𝐷𝐷 × 𝑏𝑏𝐷𝐷 + 1098.765ℎ𝐷𝐷2 − 0.115𝑏𝑏𝐷𝐷2 (7) 4. Discussions 4.1 Variance analysis of milling energy consumption model In order to analyse the relative influence between undeformed chip parameters and milling energy consumption, and to verify the accuracy of milling energy fitting model, Analysis of Variance( ANOVA) is used to analyse the milling energy consumption [26]. The variance analysis of the milling energy consumption response surface model is shown in table 5. In general, F>4 indicates that the change of design variable has a significant effect on response variables. From table 5, the F ratio of the milling energy consumption model is 75.33, it’s far larger than 4. So it can be concluded that the proposed quadratic regression model have a highly significant and the prediction accuracy is high. Meanwhile, this paper selects α =0.05 as the confidence level to determine the statistical significance of the results. When p< 0.05, it indicates that the significance of the corresponding item in the fitting model is extremely high. As is shown in table 5, the p of the fitting model is less than 0.0001, which means the statistical significance is very high, and the significance of the various dimensions in the model areℎ𝐷𝐷 , 𝑏𝑏𝐷𝐷 2 , 𝑏𝑏𝐷𝐷 , ℎ𝐷𝐷 2 and ℎ𝐷𝐷 × 𝑏𝑏𝐷𝐷 , respectively[27]. So in the milling process, un-deformed chip thickness is the main factor affecting the energy consumption
Yongbing Wang et al. / Procedia CIRP 72 (2018) 1380–1385 Yongbing Wang / Procedia CIRP 00 (2018) 000–000
1384
5
Table 3. PW6001 power analyser parameter setting Setting up the project
Connection mode
Voltage range
Current range
Current sensor mode
voltage conversion ratio （VT） 1.0
set value 3P3W3A 1~600V 1~60A 9707 Table 4. Design and result of the combined test design of surface centre Experimental design Implement of experiment Test number n/r.min-1 vf/mm/min ℎ𝐷𝐷 /mm 𝑏𝑏𝐷𝐷 /mm x1 𝑥𝑥2 1 -1.00 -1.00 0.041 2.00 2000 250 2 1.00 -1.00 0.050 2.00 2000 400 3 -1.00 1.00 0.041 4.00 2000 250 4 1.00 1.00 0.050 4.00 2000 400 5 -1.41 0.00 0.039 3.00 2000 219 6 1.41 0.00 0.052 3.00 2000 431 7 0.00 -1.41 0.046 1.59 2000 325 8 0.00 1.41 0.046 4.41 2000 325 9 0.00 0.00 0.046 3.00 2000 325 10 0.00 0.00 0.046 3.00 2000 325 11 0.00 0.00 0.046 3.00 2000 325 12 0.00 0.00 0.046 3.00 2000 325 13 0.00 0.00 0.046 3.00 2000 325 Table 5. Analysis of the variance of the energy response surface model of the milling Sum of squares of degree of Source variance of mean variance freedom Regression model 1.87 5 0.37 hD 1.74 1 1.74 bD 0.030 1 0.030 hD×bD 0.00303 1 0.00303 hD² 0.00344 1 0.00344 bD² 0.092 1 0.092 Residual Error 0.035 7 0.00497 Lack of fit 0.022 3 0.00750 Pure Errol 0.012 4 0.00307 Total 1.91 12
of milling, followed by width of uncut chip. When p>0.05, it indicates that the corresponding item in the fitting model is not significant. Therefore, the results of the analysis and table 5 indicate that the proposed quadratic regression model have a higher significance and its model can predict the milling energy consumption value of different cutting layer parameters. Meanwhile, the values of R – sq is one of the important indexes of fitting the model. It can indicate that the model is good or bad and makes a better performance of the model predicts. And when its value is close to 1, shows that the fitting quadratic response surface model prediction of the closer the value and the actual data. From table 5, the value of R - sq (0.9818) and R - sq (adjust) (0.9687) are close to the index 1. It shows that the fitting degree of milling energy consumption model is good, and predicted range of values of energy consumption is effectively. According to the experimental data and the fitting model, the normal residual distribution is shown in Fig. 5. From Fig. 5, the coincidence between predicted and measured values is high, almost straight, and the error is small. Taking the undeformed chip parameters of group 5 and group 7 as an example, the E5 and E7 was obtained as E5=3.02, E7=1.93. Compared with the measured value, Est5= 3.09, Est7= 1.99, the error is 2.27% and 3.02% respectively. And the error is within 5%, it is indicated that the experimental model and measured value fit well. Thus the model based on nonlinear quadratic regression fits the milling energy consumption prediction is very well. And the influence of input variables and response variables can be effectively mirrored by the model.
Current conversion ratio （CT） 1.0
ap/mm 2.00 2.00 4.00 4.00 3.00 3.00 1.59 4.41 3.00 3.00 3.00 3.00 3.00
ae/mm 4 4 4 4 4 4 4 4 4 4 4 4 4
Data record interval 500ms Experimental result Est/wh 2.58 1.68 2.62 1.83 3.09 1.65 1.99 2.20 2.33 2.32 2.20 2.33 2.31
F ratio
P
75.33 349.50 5.97 0.61 0.69 18.60
< 0.0001 < 0.0001 0.0446 0.4607 0.4325 0.0035
significant
2.44
0.2042
not significant
Fig.5. (a) The normal probability graph of residuals; (b) The relation diagram of the predicted value and the actual value.
4.2 Dimension reduction analysis based on milling energy consumption According to the quadratic regression equation, the corresponding surface map can be drawn. The variable factor of this experiment is the un-deformed chip thickness and the dimension reduction plane between the cutting parameters and the energy consumption of milling is obtained through the reduction of dimension. From them, the relationship between the un-deformed chip parameters ( ℎ𝐷𝐷 , 𝑏𝑏𝐷𝐷 ) and the energy response variables of the milling energy consumption is obtained. As is shown in Fig. 6.
Fig.6.Un-deformed chip thickness and width of uncut chip response surface
6
Yongbing Wang et al. / Procedia CIRP 72 (2018) 1380–1385 Yongbing Wang/ Procedia CIRP 00 (2018) 000–000
From the figure above, it can be drawn that when the undeformed chip thickness and width of uncut chip are large, the energy consumption of milling is small. When in the actual milling process, if the un-deformed chip thickness and width of uncut chip are larger, the milling time will become shorter. And meanwhile under the CNC machining conditions of stability and power under the condition of small fluctuations, it will make the machine total energy consumption small. Take the group 4, 6, 8 as example, the experimental results show that the larger un-deformed chip thickness and the larger width of uncut chip can reduce the energy consumption of the milling process. 5．Conclusion In this paper, the test design of the un-deformed chip parameters was carried out by the response surface centre combination method (CCD). And the regression model of the un-deformed chip parameters ( ℎ𝐷𝐷 , 𝑏𝑏𝐷𝐷 ) and milling energy consumption was established by using nonlinear quadratic regression equation. Based on the results above, the following conclusions can be drawn: (1)The analysis of the fitting model shows that the model can be used to analyse and predict the influence of un-deformed chip parameters on the energy consumption of milling. And the milling energy consumption decreases with the increase of the un-deformed chip thickness and the width of uncut chip. (2)When in the actual milling process, if the un-deformed chip thickness and width of uncut chip are larger, the milling time will become shorter. In order to obtain the lower milling energy consumption, it is important to select the larger undeformed chip parameters and shorten the cutting time. References [1]
International Energy Agency (IEA). Tracking industrial energy efficiency and CO2 emission [homepage on the Internet]. c2007 [updated 2007 June; cited 09.10.13]. http://www.iea.org/publications/freepublications/publication/tracking_ emissions.pdf [2] Andrea Trianni, Enrico Cagno, Stefano Farné. Barriers, drivers and decision-making process for industrial energy efficiency: A broad study among manufacturing small and medium-sized enterprises. Applied Energy.2016; 162:1537–1551. [3] International Energy Agency (IEA). International energy outlook2017. 2017 https://www.eia.gov/pressroom/presentations/mead_91417.pdf [4] Made in china 2025, 2015. http://www.agri.cn/V20/SC/jjps/201505/t20150520_4605792.htm [5] Li L, Deng XG., et al. Multi-objective optimization of tool path considering efficiency, energy-saving and carbon-emission for freeform surface milling. Journal of Cleaner Production 2017;172:33113322 [6] Liu F, Zhou XN, Cai W. The complexity and strategy for establishing product energy consumption allowance in discrete manufacturing industry. J Mech Eng 2015; 10 (19):138-45 [In Chinese]. [7] Gutowski T G, Branham M S, Dahmus J B. Thermodynamic Analysis of Resources Used in Manufacturing Processes. Environmental Science and Technology 2009; 43(5):1584–1590. [8] Pusavec F, Kramar D, Krajnik P et al. Transitioning to sustainable production – part II: evaluation of sustainable machining technologies. Journal of Cleaner Production, 2010, 18(12):1211-1221. [9] Zhao G.Y., et al. Energy Consumption in Machining: Classification, Prediction, and Reduction Strategy. Energy 2017; 133. [10] Balogun V A., P. T. Mativenga. Impact of un-deformed chip thickness on specific energy in mechanical machining processes. Journal of Cleaner Production 2014; 69.69:260-268. [11] Jia S., et al. Therblig-embedded value stream mapping method for lean energy machining. Journal of Cleaner Production;2017; 138:1081-1098 [12] J. Gawlik et al., Identification of the Minimal Thickness of Cutting
[13]
[14] [15] [16] [17]
[18] [19] [20]
[21] [22] [23] [24] [25] [26] [27]
1385
Layer Based on the Acoustic Emission Signal. Key Engineering Materials 2016; 686:39-44. Krajnik P, et al. transitioning to sustainable production – part III: developments and possibilities for integration of nanotechnology into material processing technologies. Journal of Cleaner Production 2015; 2010.1:204-204. Cai W, Liu F, Zhou X N. Fine energy consumption allowance of workpieces in the mechanical manufacturing industry. Energy 2016; 114: 622–3. Li Z Q, S. Zhang. Cutting force modeling and simulation of tee slot milling. International Conference on Mechatronics and Automation IEEE, 2009:1438-1443. Habibi Mohsen, Z. C. Chen. A semi-analytical approach to un-deformed chip boundary theory and cutting force prediction in face-hobbling of bevel gears. Computer-Aided Design 2016; 73:53-65. Zhu Kunpeng, Y. Zhang. Modeling of the instantaneous milling force per tooth with tool run-out effect in high speed ball-end milling. International Journal of Machine Tools & Manufactures 2017; 118– 119:37-48. Chen XP, He XN. Metal-cutting principles and tools. Peking University Press; Beijing; 2006 [in Chinese] Boothroyd, G., W. A. Knight. Fundamentals of Machining and Machine Tools. Marcel Dekker, 1989. Nouioua M, Yallese M A, Khettabi R, et al. Investigation of the performance of the MQL, dry, and wet turning by response surface methodology (RSM) and artificial neural network (ANN). International Journal of Advanced Manufacturing Technology 2017; (2):1-20. Fnides B, Yallese M A, Mabrouki T, et al. Application of response surface methodology for determining cutting force model in turning hardened AISI H11 hot work tool steel. Sādhanā, 2011, 36(1):109-123. Thepsonthi T, Özel T. Multi-objective process optimization for microend milling of Ti-6Al-4V titanium alloy. The International Journal of Advanced Manufacturing Technology, 2012, 63(9):903-914. Cook Dennis, Fedorov Valery. Constrained Optimization of Experimental Design. Statistics 1995; 26.2:129--178. Schudeleit Timo, S. Züst, K. Wegener. Methods for evaluation of energy efficiency of machine tools. Energy 2015; 93:1964-1970. Cai Wei, et al. A tool for assessing the energy demand and efficiency of machining systems: Energy benchmarking. Energy 2017; 138:332-347. Anderson, Marti J. A new method for non-parametric multivariate analysis of variance. Austral Ecology 2010; 26.1:32-46. Hu Luoke, et al. Minimising the machining energy consumption of a machine tool by sequencing the features of a part. Energy 2017; 121:292–305.