Prediction Model for Net Cutting Specific Energy in CNCTurning

，，，，

ZHAO Guangxi1

1.Shandong Key Laboratory of Precision Manufacturing and Non‑traditional Machining，Shandong University of Technology，Zibo 255000，P.R.China；

2.Shandong Yishui Machine Tool Factory Limited Company，Linyi 276400，P.R.China

（Received 10 July 2019；revised 27 November 2019；accepted 25 May 2020）

Abstract: A prediction model for net cutting specific energy in computer numerical control（CNC）turning based on turning parameters and tool wear is developed. The model can predict the net cutting energy consumption before turning. The prediction accuracy of the model is verified in AISI 1045 steel turning. The comparative experimental results show that the prediction accuracy of the model is significantly improved because the influence of tool wear is taken into account. Finally，the influences of turning parameters and tool wear on net cutting specific energy are studied. With the increase of cutting depth，the net cutting specific energy decreases. With the increase of spindle speed，the additional load loss power of spindle drive system increases，so the net cutting specific energy increases.The net cutting specific energy increases approximately linearly with tool wear. The results are helpful to formulate efficient and energy-saving CNC turning schemes and realize low‑carbon manufacturing.

Key words：net cutting specific energy； turning parameters； tool wear； processing scheme； low‑carbon manufacturing

0 Introduction

Manufacturing consumes electricity to convert raw materials into parts，while generating wastes and CO2emissions. The industrial sector accounts for about 31% of the total energy consumption in the USA，and the manufacturing accounts for about 60% of the energy consumption in the industrial sec‑tor. The CO2emission per unit gross domestic prod‑uct（GDP）in China is much higher than that of de‑veloped countries in Europe and America. As the world’s largest producer and consumer of machine tools，the total energy consumption of machine tools is very high in China. How to reduce energy consumption and environmental pollution in machin‑ing to achieve low-carbon manufacturing is an ur‑gent problem to solve［1-2］.

Gutowski et al. pointed out that the actual ener‑gy consumption for cutting in computer numerical control（CNC）milling，turning and grinding is less than 15% of the total energy consumption of ma‑chine tools［3］. And reasonable processing parameters are helpful to achieve energy-saving manufactur‑ing［4］. The efficient prediction models for machine tools energy consumption are of great significance for optimizing processing parameters to reduce ener‑gy consumption［5-6］. Zhang et al. divided the energy consumption of machine tools during the whole pro‑cessing into the cutting energy consumption，the aircutting energy consumption and the tool change en‑ergy consumption in the non-cutting stage，and es‑tablished the energy consumption model in each pro‑cessing stage to find processing scheme with the lowest energy consumption［7］. Specific energy is the energy consumption when removing unit volume material. Yan et al. proposed a prediction model for machine tool specific energy，which can be regarded as a function of material removal rate and spindle speed［8］. Under various constraints of lathe process‑ing，Zhou et al. developed an energy efficiency mod‑el based on multi-objective teaching and learning al‑gorithm to optimize cutting parameters［9］. Zhou et al. established an energy consumption prediction model of surface grinder based on back propagation（BP） neural network，with the grinding wheel speed，the table feed speed，and the grinding depth as input signals［10］. Liu et al. divided the input power of spindle drive system into no-load power，cutting power，and additional load loss power［11］. Jia et al.defined 15 manufacturing characteristics and energy attributes for machine tool functional components，and proposed a method for estimating energy con‑sumption in processing［12］. Aiming at the characteris‑tics of dynamic changes of processing tasks and pro‑cessing parameters，He et al. proposed a dynamic energy consumption modeling method for multi-en‑ergy sources of CNC machine tools［13］.

In summary，various energy consumption pre‑diction models of machine tool were established，and in-depth research on optimization processing pa‑rameters is conducted.However，most of the predic‑tion models do not consider the influence of tool wear on energy consumption.

The energy consumption used only to form chips and surface in cutting is called net cutting ener‑gy consumption. In hard milling，Liu et al. calculat‑ed the net cutting energy consumption based on spin‑dle power，air-cutting power，material removal rate and removal material volume［14］. Diaz et al. pro‑posed a calculation model for net cutting energy con‑sumption in milling based on cutting depth，side cut‑ting depth，feedrate，spindle speed，and removal material volume［15］.

The objectives of the paper are triple fold：（1）to establish the power and energy consumption mod‑el of CNC lathe in cutting process；（2）to develop a prediction model for net cutting specific energy in CNC turning，which is only related to turning pa‑rameters and tool wear；and（3）to reveal the influ‑ences of turning parameters and tool wear on net cut‑ting specific energy.

1 Power and Energy Consumption of CNC Lathe in Cutting

In the turning of inner and outer cylindrical sur‑face，conical surface and complex revolving surface of parts，CNC lathe consumes electric energy and converts the blank parts into products with required shape and characteristics，as shown in Fig.1. CNC lathe has many energy-consuming functional compo‑nents，including spindle system，feed axis system，cooling system，lubrication system，CNC device，lighting device，fan，tool change and other auxiliary systems［16］. The complete working process of CNC lathe includes start-up，standby，spindle start，noload operation，cutting materials，no-load opera‑tion，spindle stop，standby，and shutdown，as shown in Fig.2. In the cutting process，the power of CNC lathe includes standby powerPidle，spindle noload powerPsno，net cutting powerPtooltip，cutting flu‑id pump powerPcooland feed axis no-load power.Compared with the spindle no-load power，the feed axis no-load power is very small and can be neglect‑ed［17］. So the power of CNC lathe in cutting process can be obtained，namely

The energy consumption of CNC lathe in cut‑ting process mainly includes standby energyEidle，spindle no-load energyEsno，net cutting energyEtooltipand cutting fluid pump energyEcool.We have

Fig.1 Energy conversion in turning

Fig.2 Complete turning process

whereEidlemainly includes the energy consumption of lubrication system，numerical control device，lighting device and fan，which is related to the work‑ing time of machine tool from start to close；Esnois related to the structure of spindle drive system and the power characteristics of spindle motor，depend‑ing on the spindle speed and movement time；Ecooldepends on the power of cutting fluid pump；Etooltiprefers to the net cutting energy consumption used only for chip formation and surface generation.

2 Prediction Model for Net Cutting Specific Energy in Turning

2.1 Tool wear in cutting

The machined surface contacts and rubs with the tool flank in turning，resulting in tool flank wear［18］，as shown in Fig.3. Tool wear process can be divided into three stages：Initial wear，normal wear and sharp wear. Generally speaking，the tool wear rate is very high in the initial wear stage. In the normal wear stage，the tool wear rate is relatively low and uniform，and the wear amount increases ap‑proximately proportional to the cutting time. In the sharp wear stage，tool wear accelerates because the cutting force and the cutting temperature increase rapidly.

2.2 Prediction model for net cutting specific en⁃

ergy in turning

A method of measuring net cutting specific en‑ergy in turning by power analyzer is established in this paper. Net cutting specific energyUtooltipis de‑fined as the net cutting energy consumption when re‑moving unit volume material，that is whereQis the removal material volume.

Further，convert the net cutting specific energy into the function of the net cutting powerPtooltipand the material removal rate（MRR）.We have

According to Eq.（1），separate the standby power，spindle no-load power and cutting fluid pump power from the CNC lathe power in cutting process，then obtain the net cutting power

Substitute Eq.（5）into Eq.（4），and obtain the method of measuringUtooltipshown in Eq.（6）.

That is to say，the net cutting specific energy can be measured according to the lathe power in cut‑ting process，standby power，spindle no-load power and cutting fluid pump power. The inconvenience of this method is that the net cutting specific energy can only be measured and calculated by means of power analyzer. Establishing a simple and effective prediction model for net cutting specific energy，is of great significance for optimizing processing pa‑rameters and realizing energy-saving manufacturing.

According to metal cutting theory，the cutting power is closely related to cutting force and turning parameters. The index prediction model for net cut‑ting specific energy based on turning parameters is established

whereapis the cutting depth（unit：mm）；fthe fee‑drate（unit：mm/r）；andnthe spindle speed（unit：r/mm）.b，c，xandyare the undetermined coeffi‑cients.

Fig.3 Wear detection of tool flank VB in cutting

Turning experiments are carried out with the same workpiece material and processing parame‑ters. It is found that the net cutting specific energy varies greatly with the tool wear. This shows that the net cutting specific energy in turning is related to both turning parameters and tool wear.

On the basis of Eq.（7），an index prediction model for net cutting specific energy based on turn‑ing parameters（ap，f，n）and tool flank wear（VB）is developed in the paper

wherezis the undetermined coefficients. Then the net cutting energy consumption can be calculated ac‑cording to the removal material volume in turning

3 Experimental Verification

3.1 CNC turning experiments and power mea⁃surement

The proposed prediction model for net cutting specific energy is proved in AISI 1045 steel turn‑ing，as shown in Fig.4. The CKJ6163 CNC lathe is used，with positioning accuracy of 0.020 mm and spindle speed range of 10—1 000 r/min.

Fig.4 Turning and power measurement

The workpiece is AISI 1045 steel rod with di‑ameter of 50 mm and length of 200 mm. The CNMG120408 4025 carbide turning tool is used with cutting fluid in turning. The power and energy consumption are measured by power analyzer WT1800. The standby power of the CNC lathe is 423 W，and the curve of spindle no-load power and spindle speed is shown in Fig.5.

Fig.5 Curve of spindle no-load power and spindle speed

Sixteen groups orthogonal experiments of AISI 1045 steel cylindrical turning are carried out，and the turning parameter levels are shown in Table 1.The tool flank wear is measured twice before and af‑ter processing，and the average value is taken as the tool wear in this group of turning experiment. The turning parameters，tool wear and net cutting specif‑ic energy measured with Eq.（6）are shown in Ta‑ble 2.

Table 1 Turning parameter levels

3.2 Verification of prediction model for net cut⁃ting specific energy in turning

Firstly，the prediction model for net cutting specific energy shown in Eq.（7）is adopted，which is based on turning parameters only and regardless of the influence of tool wear. Substitute the sixteen groups of data in Table 2 into Eq.（7）to obtain an overdetermined equation group. The undetermined coefficients are calculated based on least square method：b=0.105 7，c=-0.140 6，x=0.052 3，y=0.508 0，correlation coefficientR2=0.814 3 and MSE（mean square error）=0.032 16. Then the pre‑diction model for net cutting specific energy based on turning parameters only can be obtained

Secondly，the prediction model for net cutting specific energy shown in Eq.（8）is adopted，which is based on turning parameters and tool wear. Sub‑stitute the sixteen groups of data in Table 2 into Eq.（8）to obtain an overdetermined equation group.The undetermined coefficients are calculated based on least square method：b=0.095 0，c=-0.289 4，x=0.042 4，y=0.454 3，z=2.614 4，R2=0.889 0 and MSE =0.016 71. Then the predic‑tion model for net cutting specific energy based on turning parameters and tool wear can be obtained

Thirdly，a new group of processing parameters shown in Table 3 is used to verify the accuracy of prediction model for net cutting specific energy in turning. As shown in Table 4，the net cutting specif‑ic energy measured is 2.463 7 J/mm3，the net cut‑ting specific energy predicted is 2.809 2 J/mm3with Eq（.10），and the prediction accuracy is 86.0%.While the net cutting specific energy predicted is 2.586 0 J/mm3with Eq（.11），and the prediction ac‑curacy is 95.0%.

Table 2 Turning experiment results

In conclusion，the prediction accuracy of the model shown in Eq.（11）is higher than that shown in Eq.（10）. Therefore，the influence of tool wear must be considered in the calculation and prediction of net cutting specific energy in turning.

Table 3 Processing parameters in verification experi⁃ment

Table 4 Comparison of prediction results

According to the predicted net cutting specific energy and the removal material volume during the turning，the net cutting energy consumption can be calculated with Eq.（9）. For example，in the verifi‑cation experiment shown in Table 4，the removal material volume is 20 278 mm3，the net cutting spe‑cific energy predicted with Eq.（11）is 2.586 0 J/mm3，the net cutting energy consumption predicted is 52.44 kJ while the actual net cutting energy con‑sumption is 49.96 kJ，and the predicted accuracy is 95.0%.

3.3 Influences of turning parameters and tool wear on net cutting specific energy

Turning parameters and tool wear have impor‑tant effects on cutting force and cutting temperature.Based on the prediction model for net cutting specif‑ic energy shown in Eq.（11），the influences ofap，f，nandVBon the net cutting specific energyUtooltipin AISI 1045 steel turning are studied.

As shown in Fig.6，ap，nandVBhave a great‑er influence onUtooltip. The increase ofapleads to the increase of material removal rate，soUtooltipdecreas‑es. The increase ofnresults in that the additional load loss power of spindle drive system increases，soUtooltipincreases.Utooltipincreases approximately linearly withVB. The reason is that the small edge with zero back angle is formed on the tool flank when the tool wears. As the tool wears gradually，the contact area between the tool flank and the work‑piece increases，so the cutting force andUtooltipboth increase.

In addition，fhas less influence onUtooltip.Whenfincreases，the additional load loss power of thez-axis motor of CNC lathe increases slightly.The speed and power of thez-axis motor are much smaller than that of the spindle motor，soUtooltipgrows very slowly with the increase off. For exam‑ple，theUtooltipincreases less than 0.3 J/mm3whenfincreases from 0.1 mm/r to 0.9 mm/r.

Fig.6 Influences of turning parameters and tool wear on Utooltip

4 Conclusions

The prediction model for net cutting specific en‑ergy based on turning parameters and tool wear is developed. The prediction model is proved in AISI 1045 steel turning，and the influences of turning pa‑rameters and tool wear on the net cutting specific en‑ergy are studied. The results have been used in Shandong Yishui Machine Tool Factory Limited Company，to optimize the processing parameters and improve the energy efficiency of turning scheme.The main conclusions are as follows：

（1）Tool wear has an important influence on the net cutting specific energy and the net cutting en‑ergy consumption in turning. The developed predic‑tion model for net cutting specific energy based on turning parameters and tool wear is simple，and can be used to optimize the processing parameters.

（2）In CNC turning，the increase of cutting depth leads to the increase of material removal rate，so the net cutting specific energy decreases. With the increase of the spindle speed，the additional load loss power of the spindle drive system increases，so the net cutting specific energy increases. The net cutting specific energy increases approximately lin‑early with tool wear.