CHEMICAL ENGINEERING TRANSACTIONS VOL. 59, 2017 A publication of The Italian Association of Chemical Engineering Online at www.aidic.it/cet Guest… [627790]

CHEMICAL ENGINEERING TRANSACTIONS

VOL. 59, 2017 A publication of

The Italian Association
of Chemical Engineering
Online at www.aidic.it/cet
Guest Editors: Zhuo Yan g, Junjie Ba, Jin g Pan
Copyright © 2017, AI DIC Servizi S.r.l.
ISBN 978-88-95608- 49-5 ; ISSN 2283-9216
Research on the Optimization Design of Spindle Load of CNC
Machine Tools
Xionghui Long*, Dan Su
Guangzhou Institute of Railway Technology, Guangzhou 510000, China
[anonimizat]
This paper established the finite element machine tool spindle and accessories model, accurate simulation of
the boundary conditions, constraints and external loads, and the analysis of static and dynamic characteristics
of spindle, the spindle optimizat ion design based on ANSYS software, the performance comparison before
and after optimization of the spindle; according to the results of optimizat ion design and design experience of
spindle. Analysis and interval of each size most likely made spindle, combined with orthogonal design method,
the change of the size of the spindle under the condition of static and dy namic performance analysis of the
spindle, and then obtain the optimal combination of t he size of the spindle, spindle performance comparison
optimal size composition and original spindle; based on the analysis re sults of different combinations of
spindle size the static and dynamic performance of the orthogonal ex periment, the response surface method
were obtained and the performance function related to deflection and strength and modal The performance
function, and considering the influenc e of spindle performance uncertainties , the reliability optimization design
of spindle, spindle performance anal ysis of the optimized and the original spindle; integrated optimization
design results of the comparison of t he three methods shows: considering the uncertainty of the optimization
design can according to the actual situation on the basis of improvi ng the performance of the spindle.
1. Introduction
In recent years, our country and t he developed countries in the world increasingly close ties between the
increasingly fierce competitions, the equipment ma nufacturing industry has become related to China's
international status and industrial revitalization and na tional prosperity of the important symbol (Liu et al.,
2016). Compared with foreign CNC machine tools, CNC mach ine tools in our country are reasonable in price,
relatively mature in technical specifications, and hav e certain market competitiveness (Zhu et al., 2015).
Universal milling head diagram as shown in Figure 1, Longmen CNC boring machine universal milling head is one of the core components, divided into spindle type an d mechanical spindle two categories. Electric spindle
is mainly used in high speed finishing, such as high speed rail train industry, such as aluminum alloy, titanium
alloy, high temperature resistant a lloy and composite material, such as high-strength lightweight parts.
Mechanical spindle type is mainly used in high torque st rong cutting, such as the energy industry of the turbine
blade, the ship industry, such as screw propeller parts (Zhou and He, 2015). Longmen boring machine spindle type universal milling head has two mutually perpendicu lar planes of rotation, not only can complete the
milling work in machining surface, can be obtained through various spindle position universal milling head horizontal and oblique axes of rotation, in order to ad apt to the different requirements of the edges, grooves
and other processing various parts in particular, it has important practical value in the two maintenances.
Longmen GMC2000 boring machine universal milling head with double swing head is the appropriate speed,
dynamic angle, An axis swing angle of plus or minus 110 degrees, C axis swing angle + 360 degrees. The tool
center is always maintained in the normal direction of t he work piece, and avoids zero cutting speed. Electric
spindle rotary tool, high speed (up to 24000rpm), especially suitable for processing mound and non-ferrous
metal materials, can also be different spindle power an d torque configured according to user requirements,
CNC machine tool structure schematic diagram shown in Figure 1.

DOI: 10.3303/CET1759012

Please cite this article as: Xionghui Long, Dan Su, 2017, Research on the optimization design of spindle load of cnc machine to ols, Chemical
Engineering Transactions, 59, 67 -72 DOI:10.3303/CET1759012
67

Figure 1: Schematic diagram of universal milling head and Schematic diagram of the structure of NC chine
tools
The spindle is the core component of the universal milli ng head, whose function is to drive the cutting tool
(grinding wheel) to rotate. The spindle has a great influence on the machining accuracy, surface quality and
productivity. With the development of modern industry to the requirement of machine machining precision and
efficiency increasing requirements on the performance of machine tools spindle is higher: wide speed range,
high precision, high rigidity, small vibration, low no ise, small deformation, and has good resistance ability of
forced vibration and self-exc ited vibration. (Wan, 2014).
2. Literature review and theoretical exposition
2.1 Introduction to finite element method
Finite element method (FEM) is a kind of highly effe ctive and commonly used numerical method. The field of
scientific computing, often need to solve all kinds of differential equations, and many analytical solutions of
differential equations is generally difficult to obtain, us ing the finite element method of differential equation is
discretized, can be programmed using computer aided solution. The solution domain discrete continuous
combined body of a unit, the approximate function hypothes is in each unit to slice representation of the
unknown field function solution domain to solve, approx imate function is usually expressed by the unknown
function and its derivative numerical interpolation f unction at each node unit (Zhang et al., 2014). Thus, a
continuous infinite degree of freedom problem is transformed into a discrete finite degree of freedom problem. Feng Kang summed up this, made systematic theoretical re sults (Liu et al., 2015). The finite element method
solving domain is discretized into finite element (grid), unit node connection between nodes, transfer
interaction between units, node field as a function of the unknown quantity, and piecewise difference field
distribution simulation function, and solving. Therefor e, the complicated engineeri ng problem is reduced to a
simple problem with high accuracy, and the stress and stra in distribution of the element is shown in Figure 2.

Figure 2: Schematic diagram of el ement stress and strain distribution
2.2 Basic theory of orthogonal test design
In order to determine the optimal comb ination of factors, the authors put forward a kind of fast and efficient
experimental design method, orthogonal test design (Duan and Ding, 2015). The orthogonal test design is a
design method of optimal level for multi factor experiment, which is bas ed on the orthogonalit y of the selected
part of the comprehensive test representative, the test of representative and comprehensive test compared
with the equilibrium dispersion char acteristics of neat comparable". The orthogonal test design is a scientific
68

method for the design and analysis of a proper orthogonal table. The orthogonal design is a design method of
multi factors and multi levels, it is based on the orthogonality selected from comprehensive test of some
representative points test, and these re presentative points have "evenly disp ersed, neat features than the
orthogonal test design is the main method of fractional factorial design. Suppose that a comprehensive test of
the three factors, each of which has three level tests, then we need 33 =27 test, comprehensive experiments
of three factors and three levels of t he diagram as shown in Figure 3, in th e orthogonal test, each level of each
factor and each factor collocation are un iform, only three times the test. So the orthogonal test can get almost
the same effect as the full-scale experiment. In the three combinations, the level of factor B and factor C is
only one of the three levels of factor A, which is the same as the. That is to say, if we compare three influence
factors of A level of the test result s, the experimental groups containing A1, A2, A3 were added, B and C for
the factors of A, equivalent to a fixed.

Figure 3: Schematic diagram of comprehensive experiments of three factors and three levels
2.3 Orthogonal experiment design
The orthogonal test design includes: designing ortho gonal test scheme, orthogo nal test, calculation and
analysis of the experimental results and the four parts of the test. The first step is to clarify the purpose of the
design. In the orthogonal desi gn, the first step is to clarify what is the purpose of the test, or what is the reason,
which is the basis of orthogonal test. After dete rmining the purpose of the ex periment, then determine the
corresponding test indicators. The test index is a measur e of the value of the experimental results. Secondly,
choose the appropriate factors and fa ctors (Wang et al., 2016). The first step and the second step are very
dependent on the experimental design experience and expe rtise of the experimenter. A good experimenter
can successfully complete the orthogonal test, and get good results. Secondly, c hoose the orthogonal table
and table design. The reliability optimization design considering random variables and interval variables can
obtain the optimal combinatio n of design variables under the condition of ensuring the reliability of the results.
In this section, we mainly introduce the reliability opt imization design of double loop, which is the reliability
optimization design under the condition of random variables and interval variables (Yang, 2016). The results of
the orthogonal test results are show n in Figure 4, and the core codes are listed as follows (Li and Zhang,
2016).

Figure 4: Schematic diagram of digital building technology
1. settings = cordexch(3,13,'q'); % settings; 2. mr = p(2,:); % mr: middle row, i.e., middle level 3. mr =
mr(ones(13,1),:); 4. hr = (p(1,:) – p( 3,:))/2; 5. hr = hr(ones(13,1),:); 6. expCond = settings.*hr + mr; 7. p1 =
expCond(:,1); 8. p2 = expCond(:,2); 9. p3 = expCond(:,3); 10. y = zeros(13,1); 11. for k = 1:13; 12. y(k) =
69

1.25*(p2(k) – p3(k)/1.5183)./(1 + 0.064*p1(k) … ; 13. + 0.0378*p2(k) + 0.1326*p3(k))* normrnd(1,0.02); 14. end;
15. data = [expCond y]; 16. settings = cordexch(3,13,'q') ; % settings; 17. mr = p(2,:); % mr: middle row, i.e.,
middle level; 18. mr = mr(ones(13,1),:); 19. hr = (p(1,:) – p(3,:))/2; 20. hr = hr( ones(13,1),:); 21. expCond =
settings.*hr + mr; 22. p1 = expCond(:,1); 23. p2 = expCo nd(:,2); 24. p3 = expCond(:,3); 25. y = zeros(13,1);
26. for k = 1:13; 27. y(k) = 1.25*(p2(k) – p3( k)/1.5183)./(1 + 0.064*p1(k) …; 28. + 0.0378*p2(k) +
0.1326*p3(k))*normrnd(1,0.02); 29. end; 30. data = [expCond y]
3. Finite element analysis and optimization design of spindle
3.1 Analysis and calculation of spindle and bearing
The main content of this chapter is the quantitative anal ysis of the external forces of the spindle components,
simplifying the combination of bearing parts and bearings and spindle, the establishment of accurate digital
simulation model, in order to analyze the static and dynamic characteristics of the spindle. Finally, the optimal
design based on ANSYS software. In order to establish the finite element model of the spindle is simple and
effective, we must thoroughly under stand the Longmen boring machine univ ersal milling head structure and its
operation mode. The cutting force decomposition is shown in Figure 5, the static and dynamic problems of the
shaft, according to the combinat ion between the bearing and the sha ft and the bearing surface modeling
problem, and the nonlinear contact problem of the bearing itse lf is not the focus, simplify the combination
between the bearing and the shaft and the bearing surface does not affect the purpose of the study, but also
improve the modeling speed and precis ion, and improve computing effici ency. Therefore, the bearing is
simplified as a simple radial spring and damper characteristics by using the finite difference analysis model of
Longmen boring machine spindle. The main intention of def ormation icon as shown in Figure 5, in combination
with the actual simulation of beari ng and the bearing and the shaft surfac e, the node consolid ation of the end
of the spring on the main shaft, and the other end (P1, P2, P3, P4 four nodes) for complete consolidation,
each radial stiffness (K/2 single row radial thrust ball bearing stiffness). The 3 column ball bearing front spindle
with radial 3 columns of the same radial spring distributi on to simulate the bearing st iffness, axial column using
3 axial spring simulation bearing stiffness; simulation above 2 column ball bearing spindle end. In the analysis
and design of the spindle system, the bearing stiffness is one of the important factors, but due to the nonlinear
bearing stiffness and bearing type, material, shape and other factors, so to accurately determine the bearing
stiffness is not realistic. Zhu et al ., the use of inductance measurement of a set of bearing stiffness test device,
more accurate measurement of the stiffness of the bearing (Ma et al., 2014).

Figure 5: Schematic diagram of cutting force decomposition and Schematic diagram of spindle deformation
3.2 Spindle finite element analysis
Finite element analysis is the foundat ion of the whole research. It includes the selection of unit type, the
establishment of numerical analysis mo del and finite element analysis. Determine the type of units required for
the network before the model is est ablished. The selection of unit type is based on the analysis of the
complexity of the object, the characteristics of t he stress and strain, the accuracy of the calculation
requirements and the computing power of the computer. According to the characteristics of the solid modeling
of the machine tool spindle system a nd the structural mechanical characte ristics reflected in the motion, the
SOLID45 unit is used to divide the machine tool spindle components. The SOLID 45 element is used for 3D
solid structure model for 3-D solid structure unit. T he unit consists of eight nodes. Each node has three
degrees of freedom in X, y, and Z, as shown in Figu re 6 represents a deformed SO LID45 unit. SOLID45 unit
can be simple control, more accurate simulation of the actual situation. The SOLI D45 element can be used to
simulate plasticity, creep and expansion, but also to meet the stress hardening, large deformation and large
strain characteristics. In order to simplify the me shing of the bearing with COMBIN14 element, the element
70

has the axial or torsional elastic damping properties of one-dimensional, two-dimensional or three-dimensional.
The model is the basis of network, has a direct impact on the process of network model is reasonable. If the
model is large and regular, ANSYS software provides a method for static analysis and dynamic analysis. In
addition, the distributed comp uting method, the bandwidth optimization and the wave front processing method.
Longmen GMC2000 boring machine spindle finite element model as shown in Figure 6, the internal model as
shown in figure 6. In the model of the Z coordinates of the 126.5 small ladders, t he analysis can be ignored. In
order to more accurately simulate the axis in the work fo rce, the axial force of the work directly loaded in the
handle, so the handle part modeling as a part of the spin dle. When the boundary condition is loaded, the rear
end of the spindle does not change when the motion is carried out, so the rear end of the model spindle is
completely fixed. Because the front-end milling cutter, so draw a solid. Finite element analysis is based on the
solid grid. If the calculation precision of the result is high, the precision of meshing should be improved.
Otherwise, the result of finite element analysis does not accord with the actual requirement, which will
seriously affect the effect of the whole research projec t. If the structure of the research object is complex, the
structure of the key part of the refi nement of the grid, for example, the temperature gradient, stress gradient,
large deformation of the structure, in the non-critical part of the coarse mesh. The purpose of model checking
is to check whether the mesh is reasonable. Howeve r, because the model is not the geometry of the ideal
rules, the actual grid cannot be a rule. So it is necessary to check the mode l in order to limit the deformation of
the mesh to a certain extent.

Figure 6: Schematic diagram of SOLI D45 unit structure and Schematic diagram of the spindle structure after
loading of spindle 1/4 boundary conditions and Schematic diagram
4. Orthogonal test design and analysis of spindle load
In this paper, we choose the comprehensive balance met hod to analyze the results of orthogonal test. The
effects of different factors on the test results were different. The larger the range, indicating that the change of
the factors in the test range will lead to the change of the test index, so the line is t he biggest influence on the
test results, which is the most im portant factor. To determi ne the optimal scheme, such as the index of the
bigger the better, intuitive sake, draw the factors and indicators of the trend, the trend diagram as shown in
figure 7, select the optimal scheme of comprehensive balance: 1 factors, 5 indexes are 3 level, the choice of 3
levels.

Figure 7: Schematic diagram of the trend factor
71

With the development of modern com puter technology, and in-depth stud y of response surface method and
improvement of the theory, the scientific computing m odel is becoming more and more complicated, the
response surface method is obtained including chemistr y, and biology and engineering favor (Hui et al., 2014).
It has been proved that the response surface method has a high status in the field of engineering, such as
reliability analysis, optimization design, dynamics research and engineering process control. In order to apply
the two-order response surface approximation model, the first step is to find the optimal region by using the
low order method, and then the polynomial approximation is used in the optimal region.
5. Conclusions
The importance of machine tool spindle is self-evid ent, many scholars devoted their lives to improve the
performance of the machine tool spindle, and pay it to practice. Based on the optimization design of the
spindle of the spindle, reduce weight, reduce materi al consumption and production cost, improve the working
performance of the machine, can great ly shorten the design cycle of t he spindle, and obtained the most
consistent with the actual situation of the optimization method. An accurate finite element of machine tool
spindle parts and accessories of the model, analysis of spindle of t he static and dynamic performance,
optimized design based on ANSYS software, and compared the performance of the spindle before and after
optimization; response surface method were obtained and the performance function related to deflection,
strength and modal based on the perfo rmance of the resulting function, comprehensive considering the
influence of spindle performance uncertainties, the reliability optimization design of spindle and spindle after
optimization and performance analysis of the spindle, and comprehensiv e optimization design results of the
comparison of the three methods, the most practical method can great ly improve the spindle safety
performance and processing performance.
Acknowledgments
This work is supported by the research and practice of Guangdong Province vocational education numerical
control skill competition information platform construction (Guangdong educat ion and scientific research "11th
Five-Year" planning 2011 Annual project) (2011TJK252).
Reference
An Huping., Zhang Z., Yang T., 2014, Study on the desig n method of machine tool spindle parameters based
on elastic deformation, Mechanical m anufacture and automation, 5, 9-12.
Duan P., Ding X., 2015, The lightweight design of mach ine moving parts based on the theory of equivalent
static load, Journal of University of Shangh ai for Science and Technology, 37(6), 583-588.
Hui Y., Huang Y., Yang X., 2014, Since a full load turning center stiffness of bolted joint detection and design
application, Journal of mechanic al engineering, 50(17), 142-148.
Li R., Zhang L., 2016, The study on the estimation method of NC machine tool load stability collision
probability design, Bulletin of science and technology, 32, 11,
Liu K., Luo W., Lai Y., 2015, Study on Optimization Design of machine tool spindle subspace method based
on improved parallel, Electrical and mechanical engineering, 32(6), 788-792.
Liu S., Zhang Y., Wang H., 2016, Research on expert syst em of optimization of NC machine tool spindle,
Journal of agricultural machinery, 47(4), 372-381.
Ma Y., Yue C., Hao S., 2014, Finite element analysis of t he spindle of special machine tools of dark straight
European wooden tenon, Journal of Northwest Forestry University. 29(1), 185-187.
Wan H., 2014, Vibration Analysis for the Spindle of Ma chine Tools under Cutting Loads, Modular machine tool
and automatic manufacturing technique, 4, 22-25.
Wang C., Wang Y., Zou S., Peng L., Hu J., Yao Q., 2016, Modal analysis of high precision servo cnc camshaft
grinding machine's grinding carriage, Machine Tool & Hydraulics, 3, 79-83.
Yang P., 2016, Radiation-based virus attack and defense re liability optimization design, Chemical Engineering
Transactions, 51, 793-798, DOI: 10.3303/CET1651133.
Zhang S., Ma W., Li J., 2014, Study on the contact char acteristics of high speed CNC machine tool spindle
angular contact ball bearin g, The bearing, 9, 1-5.
Zhou S., He X., 2015, Inch Hwayoung, spindle optimizati on design based on reliability analysis, Machine tool
& hydraulics, 43(5), 164-166.
Zhu C., Nobu H., Li T., 2015, Optimization design of CNC machine tool spindle based on improved PSO
algorithm, Chinese mechanical engineering, 26(20), 2784-2788.

72