Collaborative manufacturing technologies of structure shape and surface integrity for complex thin-walled components of aero-engine: Status,challenge and tendency

Bio ZHAO, Wnfng DING,*, Zhong SHAN, Jun WANG,Chngfng YAO, Zhngi ZHAO, Ji LIU, Shihong XIAO, Yu DING,Xiowi TANG, Xingho WANG, Yufng WANG, Xin WANG

a College of Mechanical and Electrical Engineering, Nanjing University of Aeronautics and Astronautics, Nanjing 210016, China

b School of Mechanical Engineering, Northwestern Polytechnical University, Xi’an 710072, China

c AVIC Manufacturing Technology Institute, Beijing 100024, China

d School of Mechanical Science and Engineering, Huazhong University of Science and Technology, Wuhan 430074, China

e AECC Shenyang Liming Aero-engine Co., LTD., Shenyang 110000, China

KEYWORDS Collaborative manufacturing of shape and performance;Complex thin-walled component;Intelligent process control;Material removal mechanism;Surface integrity

Abstract Presently, the service performance of new-generation high-tech equipment is directly affected by the manufacturing quality of complex thin-walled components.A high-efficiency and quality manufacturing of these complex thin-walled components creates a bottleneck that needs to be solved urgently in machinery manufacturing.To address this problem,the collaborative manufacturing of structure shape and surface integrity has emerged as a new process that can shorten processing cycles, improve machining qualities, and reduce costs.This paper summarises the research status on the material removal mechanism,precision control of structure shape,machined surface integrity control and intelligent process control technology of complex thin-walled components.Numerous solutions and technical approaches are then put forward to solve the critical problems in the high-performance manufacturing of complex thin-wall components.The development status, challenge and tendency of collaborative manufacturing technologies in the high-efficiency and quality manufacturing of complex thin-wall components is also discussed.

1.Introduction

Complex thin-walled components have demonstrated broad application prospects in aerospace, nuclear reaction and vehicles owing to their superior comprehensive properties, such as their light weight and high aerodynamic efficiency (Fig.1).1–6In aerospace, the requirements for these components (e.g.blisk, casing and turbine shafts) are constantly increasing due to the rapid development of aero-engines towards strict service conditions (e.g.higher temperature, pressure and speed)and service demands(e.g.longer service life and greater reliability), as shown in Fig.2.7–10For example, the America has widely adopted integral disc structures in the compressors of 4G aero-engines (F119) to improve their service performance compared with 3G aero-engines (F100) in terms of thrust (100 % improvement), durability (2 × improvement)and engine life (1.5 × improvement).Meanwhile, complex thin-walled components have also been employed in 4G/5G military engines and civil engines with high bypass ratio.However, some problems emerge after the formation of a complex structure blank, including inaccurate benchmarks, uneven distribution of allowances, complex geometric characteristics,complex multi-field coupling mechanism of time-varying processes, large material removal volume, poor rigidity, dynamic cutting with high sensitivity,easy elastic deformation and chatter and extreme difficulty in controlling manufacturing consistency and accuracy11–15.

Fig.1 Complex thin-walled components utilized in various fields1–6.

The excess blank materials of complex thin-walled components should be removed using a suitable method owing to the restrictions of near-net forming technology and economic affordability.Accordingly, the main manufacturing methods,including mechanical energy field machining (e.g.cutting and grinding), electrical energy field machining (e.g.electrolysis,electroforming and laser machining), composite energy field machining (e.g.ultrasonic vibration-assisted cutting, laserassisted electromachining and laser heating-assisted cutting),play indispensable roles in manufacturing the key components of aero-engines.16–19However, the coupling effect of the mechanical/electrical/composite energy field on the force, heat and flow field of material removal is prone to produce different degrees of damage on the newly machined surface layer of components.These types of damage(e.g.severe residual tensile stress, strain hardening, micro-cracks and white layers)can be ascribed to the poor integrity of the machined surface, which leads to the service performance of components,especially fatigue resistance, being much lower than that of the theoretical service performance of raw materials.20–24In this case,machined surface integrity control and shape precision control are equally important during the manufacturing process given their decisive impacts on the service performance of aeroengine components.In addition, the key components (e.g.blisk and thin-walled casing) of high thrust-to-weight ratio aero-engines,are mainly manufactured by integral cutting processes with forged blanks, such as cutting and electrolysis.However, the large cutting force, high temperature, and the fast tool wear,as well as the strong time-varying characteristics appear in machining these difficult-to-cut materials.25,26Moreover,the machining system has more prominent characteristics of weak rigidity due to the complex surface and thinner wall thickness structures, and thus the deformation and vibration are prone to happen, resulting in the poor machining stability.27In this case, the extreme challenge resulted from the two constraints of geometric accuracy control and stability control can bring about complex coupling effect of uncertainty during machining process.Then, the severe surface integrity control problems have to be faced, such as the influences of the high-efficiency milling and low-stress electrolysis on surface state of subsequent precision machining, as well as the accessibility of complex thin-walled structures on the quality uniformity and consistency of the subsequent machined surface modification and strengthening process.The abovementioned challenges restrict the further promotion on the high reliability and long service life of engines, resulting from the unsatisfied collaborative manufacturing level of structural shape and surface integrity of key components.Along this line of consideration, the collaborative manufacturing for structural shape and surface integrity of key components is a matter of great urgency.

A basic research system on the manufacturing technologies of aero-engine components has not yet been established given the long-standing dominance of imitation and catch-up in China.In addition, a profound understanding of the ‘collaborative control mechanism of structure shape and surface integrity for manufacturing of aero-engine components’in nuclear science remains lacking.A scientific basis and a reliable method for developing collaborative manufacturing technology strategies of shape and performance have also been absent for a long time, and the available process parameters mainly depend on trial-and-error method or refer to existed process trials.Furthermore, the basic theories and key technologies for the collaborative manufacturing technologies of structure shape and surface integrity for complex thin-walled components are severely lacking, thereby leading to a bottleneck in the high-performance manufacturing of these components.In recent years, scholars have started to explore the basic theory and applied technology of collaborative manufacturing technologies in foreign leading aero-engine enterprises (e.g.American General Electric, Pratt & Whitney and British Rolls-Royce).The collaborative controls of structure shape and surface integrity have been primarily realized through intelligent manufacturing methods, thus significantly improving the service performance and life of aero-engines.As can be seen in Fig.3, the number of related literature on material removal mechanism,machining accuracy control,surface integrity control and intelligent manufacturing process control gradually increased from 43 to over 500 over the past 20 years.The number of studies on machining accuracy control reported the fastest increase over the recent five years, whereas that on other topics reported varying degrees of increase (Fig.3(a)).The number of papers on collaborative manufacturing technologies in aerospace published in China is also increasing (Fig.3(b)),and even exceeded the number of papers published in America after 2011.This phenomenon only underscores the increasing research attention on manufacturing technologies in the aerospace field, which has become a development hotspot and research frontier in the field of advanced manufacturing technology.However, studies on the collaborative control technologies of structure shape and surface integrity for the manufacturing of aero-engine components are still lacking.

This paper explores the progress of research on the collaborative manufacturing technologies of structure shape and surface integrity for complex thin-walled components, material removal mechanism, precision control of structure shape,machined surface integrity control,and intelligent process control technology at home and abroad.The key problems and associated solutions in the manufacturing of complex thinwalled components are also discussed.Some directions for future research on the high-performance manufacturing technologies for complex thin-walled components are also proposed.Fig.4 presents a flowchart of the development of collaborative manufacturing technologies of structure shape and surface integrity for complex thin-walled components.

2.Material removal mechanism

Fig.2 Aero-engine and the associated typical complex thin-walled components7–10.

Fig.3 Number of collected papers in SCI on manufacturing technologies used in aerospace.

Fig.4 Collaborative manufacturing technologies of structure shape and surface integrity for complex thin-walled components.

Certain manufacturing processes, such as cutting, electrical machining (e.g.electrolysis) and composite energy field machining (e.g.ultrasonic vibration-assisted cutting), have been widely adopted recently to remove the materials of aeroengine components.28–30The interactions of the mechanical energy field, electrical energy field or composite energy field with difficult-to-cut materials are used to remove these materials and produce a workpiece surface.How the material removal mechanism serves as an important scientific basis for controlling the shape accuracy and surface integrity of complex thin-walled components has always received increasing attention and has eventually become a research focus in the advanced manufacturing of aero-engines31–33.

2.1.Cutting technology and associated material removal mechanism

Cutting technology is considered the main method in machining the blisk or thin-walled casing of aero-engines due to its extraordinary properties, including high stability, flexibility and precision.34,35In the late 1970 s, General Electric applied a five-coordinate computer numerical control (CNC) milling technology in the manufacturing of machine aero-engine(T700) blades and developed the associated machine tool systems.Subsequently,a five-coordinate CNC milling production line was established in other world-renowned aero-engine manufacturers, including Pratt & Whitney, Rolls-Royce and MTU, for machining blades and blisks.NR Electric then developed an electrical MAX-CAM software system for machining blisks,which remarkably improved their machining efficiency and quality.Since then, the manufacturing of complex thin-walled components became mainstream.

Recent years have witnessed significant advancements in the material removal and surface formation mechanisms of high-temperature resistance materials for aero-engine components.Zhang et al.36found that the closed high-strain strength of adiabatic shear bands could predict the fractures formation during the cutting of titanium alloys.Apart from strain strength, the fracture process can be controlled by altering the critical bearing capacity and critical failure strain to realize material removal.Wang et al.37summarized the previous research on the material removal mechanism in metal highspeed cutting and found that the change in material dynamic properties is the fundamental driver behind the changes in chip morphology and material removal mechanism.Harzallah et al.38revealed that the sawtooth-shaped chip of titanium alloys is synthetically affected by the adiabatic shear and crack propagation of materials.The crack propagation initially appeared at the tool tip and then extended to the free surface in the shear zone.The formation of these periodic sawtooth chips eventually led to the periodic vibration of cutting forces.They also employed a thermal infrared imager to measure the strain, strain rate, temperature and dissipated power along with the variations in displacement, velocity and crack propagation to further understand the material removal mechanism.A sawtooth chip or fractured chip tended to form due to the local strain, especially at high cutting speeds.A 3D finite element orthogonal cutting model and the associated analysis results were then presented to reveal the material removal mechanism as shown in Fig.5.Sonawane and Joshi39found that the width and thickness of undeformed chips increased by 62 % and 39 %, respectively, as the inclination angle of Inconel718 superalloys varied from 0° to 45°.Wang et al.40revealed that when the cutting speed reached 7000 m/min,the chip morphology of Inconel718 superalloys is transformed from sawtooth to fragmentary as illustrated in Fig.6(a).In addition, the localization of shearing chips is significantly affected by the initial yield stress and thermal softening coefficient.Ullah et al.41investigated the influence of the thermomechanical effect on the forming mechanism of sawtooth chips in high-speed milling titanium alloys (Fig.6(b)) and verified their numerical results via experiments.Zhang et al.42revealed the deformation mechanism and mechanical behavior of the workpiece material in the chip-forming zone of abrasive grains by studying the changes of output parameters of high-speed grinding with different lubrication conditions.When the lubrication conditions in the material removal process are different,it would change the strength relationship between the strain rate strengthening effect and the thermal softening effect,thus changing the material removal mechanism43–45.

Fig.5 Geometry and boundary conditions used for the implementation of a 3D FE orthogonal cutting model38.

Fig.6 Evolution of chip morphology at various cutting speeds.

Fig.7 Material removal mechanism in MECM53.

Scholars have conducted extensive and in-depth research on the removal mechanism of material cutting.However, the current machining research still focuses on blocks or simple structures, which fails to consider the influence of complex structural geometry factors on the material removal mechanism.Moreover,there are lacking theoretical and experimental researches on the cutting of complex thin-walled structures.

2.2.Electrochemical machining technology and material removal mechanism

Electrochemical machining (ECM) as a non-contact machining method that removes materials via electrochemical reaction without causing tool wear, surface recasting layer and residual stress regardless of the mechanical properties (e.g.strength and hardness) of workpiece materials.46,47ECM has emerged as a main manufacturing technology for key components, such as blades and blisks.48,49MTU Corporation in Germany regarded ECM as its preferred manufacturing technology for machining difficult-to-cut materials, such as blades and blisks.Meanwhile, Sermatech Corporation in America also adopted ECM for machining the important components of aero-engines.Rolls-Royce in the UK developed an electrolytic machine tool for blisks and established a production line.According to the MTU, the milling and ECM processes each accounted for 45 % of all blisk manufacturing technologies in the world.Recent years have witnessed remarkable achievements in high-efficiency and precision machining,machining gap modelling, high-temperature material dissolution behaviour, electrolyte flow field optimization and formation mechanism of surface defects.50–52Wang and Qu53investigated in depth the material removal mechanism of mechano-electrochemical milling (MECM) and reported three material removal processes involved in this mechanism.Fig.7 presents the experimental setup and physical models.

Machining efficiency and low-cost advantages continued to emerge along with the development of native electrolysis technologies for machine blisks.The integrated process with the rough machining method using electrolysis and the subsequent finish machining method by milling has gradually become an important tendency and attracted great concerns in machining blisk, as illustrated in Fig.8.A stable and efficient machining of the blisk cascade channel is realized by using the processing methods (e.g.radial feed machining, nesting machining and tube electrode CNC machining) proposed by domestic research institutes, including the Nanjing University of Aeronautics and Astronautics and the Aeronautical Manufacturing Technology Research Institute of China.54–56However, the influence of redistribution of stress caused by the connected processes between milling and electrolysis machining on the deformation of weak rigid blades has been largely ignored.In addition, the influence of surface state reconstruction caused by the electrical and mechanical energy field on the fatigue properties of blade is still unknown.In this case, electrochemical dissolution and material removal mechanisms need to be developed, and the residual stress distribution, surface formation and reconstruction rule under different processing procedures should be revealed.The design and control method of blank allowance should also be controlled to subsequently design an integrated process that can support a highefficiency and accurate manufacturing of blisks.

At present,the influences of the stress redistribution on the deformation of weak rigid blades has been largely ignored during the successive milling and electrolysis processes.Moreover,the material removal mechanism of complex thin-wall structures is not clear currently under the milling and electrolysis energy field coupling effects.In the further study, the simulation and experimental researches on machining process need to be conducted with the consideration of complex geometric structures and electrolytic process characteristics, aiming at clarifying the multi-field coupling mechanism and material dissolution law.Furthermore, a basis for the selection of process parameters and electrolyte parameters in the subsequent electrolytic low-stress machining will be then provided.

2.3.Ultrasonic-assisted cutting technology and material removal mechanism

As an important composite energy field processing technology,ultrasonic-vibration-assisted cutting (UVAC) has witnessed a rapid growth in machining difficult-to-cut materials being used in the aerospace field.57,58In this technology,the regular ultrasonic frequency vibration with one or several controllable vibrating directions is applied on cutting tools or workpieces during conventional cutting processes, thereby transforming the material removal behaviour from continuous cutting to periodic discontinuous cutting as illustrated in Fig.9.59Compared with the conventional cutting process, the influence of the force-thermal coupling can be effectively reduced in UVAC, thereby resulting in variations in the chip removal behaviour.

Fig.8 Integrated machining procedures for aero-engine blisks.

Fig.9 Schematic diagram of UVAC processes59.

Numerous studies have investigated UVAC and the material removal mechanism recently.For instance, Tong et al.60employed elliptical ultrasonic vibration-assisted machining technology to reveal the surface morphology of aluminum alloys and obtained the better machining quality using the appropriate amplitude and low feed rate.Gao et al.61conducted the longitudinal-torsion ultrasonic vibration assisted machining trials with Ti–6Al–4 V alloys, and then presented that the lower milling forces and longer service life were achieved compared to the traditional machining method.Liu et al.62studied the cutting behavior with in-situ TiB2/7050Al metal matrix composites using an axial ultrasonic vibration milling method.Their pointed out that the milling force could be significantly reduced and the tool life was increased by more than 2 times once employing axial ultrasonic vibrations into traditional milling processes.Peng et al.63studied the dynamic cutting thickness of the ultrasonic turning of thin-walled titanium alloy cylinders based on established critical cutting thickness models, improved the critical cutting thickness via ultrasonic machining and found that the ultrasonic vibration benefits the chip breaking and machining stability.Babitsky et al.64found that the machined surface roughness can be reduced by 50%by adopting ultrasonic vibration in the turning processes of Inconel718 superalloys.They explained that the high-frequent impact increased the whole stiffness of the machine tool-cutting tool-workpiece system when using ultrasonic vibration.In addition, the generated built-up edges on tools can be eliminated by ultrasonic vibration due to the reduced machined surface roughness.Gao et al.65systematically reviewed the influence mechanism of ultrasonic vibration-assisted machining on the milling and grinding surface quality of CFRP.They concluded that ultrasonic vibration could reduce the surface roughness by about 30 %–40%in most cases.Patil et al.66reported that the cutting temperature remarkably decreases by 48%in machining titanium alloys when ultrasonic vibration is applied due to the reduced contacting time with cutting tools and chips and the improved convection cooling of cutting tools.Chen et al.67found that the tool wear can be decreased by 30 % due to the separation properties of the tool-workpiece system in the UVAC of titanium alloys and that the formation of burrs can be restrained.Ni and Zhu68studied the cooling and lubricating mechanism(Fig.10) affected by ultrasonic vibrations in cutting titanium alloys and revealed that the cutting force can be reduced by 30 % to 55 %.

Numerous investigations focus on various ultrasonic vibration-assisted methods (e.g.one-dimensional and twodimensional ultrasonic vibrations) and material removal mechanisms during machining difficult-to-cut materials.However, the force-heat-structure coupling effect and the associated material removal behavior in ultrasonic cutting complex thin-walled components are not clear, restricting the application of the ultrasonic cutting technology.In sum, the material removal and surface formation mechanisms during the cutting (mechanical energy field), electrolysis (electrical energy field), ultrasonic cutting (composite energy field) and other machining processes have been stated recently.Here, these studies have mainly focused on the workpiece with simple structures, whereas the further investigations on key components (e.g.blisks, thin-walled casings,etc.) with complex structures should be performed.In addition, the available material removal theories still lack to achieve the high efficiency and quality machining of complex thin-walled components, which should be enhanced to reveal material removal and surface formation mechanisms.Moreover, the existed machining theory should also be integrated with corresponding theories for collaborative manufacturing technologies of structure shape and surface integrity of the key components in aero-engines.

Fig.10 Material removal mechanism of UVAC68.

3.Machining accuracy control

Numerous problems need to be addressed after the formation of a complex structure blank, such as inaccurate benchmarks,uneven distribution of allowances, complex geometric characteristics, complex multi-field coupling mechanism of timevarying processes, large material removal volume, poor rigidity, dynamic cutting with high sensitivity, easy deformation and chatter and extreme difficulties in controlling manufacturing consistency and accuracy.69,70Researchers at home and abroad have focused on the cutting deformation prediction and control of complex thin-walled components, the dynamic response and chatter suppression of time-varying process systems and the adaptive machining and error control of complex thin-walled components71–74.

3.1.Cutting deformation prediction and control of complex thinwalled components

The elastic–plastic and residual stress deformation can affect the machining accuracy during the cutting processes of complex thin-walled components.75–79Akhtar et al.80established a deformation prediction model in the milling of titanium alloys by using elastoplastic deformation numerical simulation whilst considering the effect of force-thermal coupling.Fontaine et al.81predicted workpiece deformation by using the cutting force analytical model in the milling process and found that the cutting deformation is mainly affected by the cutting conditions, tool runout and workpiece inclination angle.

Schulze et al.82established a dynamic prediction model of the residual stress under the influence of the milling process on residual stress redistribution of thin-walled components.Yao et al.83applied the delamination discretisation of residual stress in predicting machining deformation.Tang et al.84built an elastic–plastic deformation prediction model that involves the coupling of the initial residual stress and machining load.Gao et al.85established an analytical prediction model for the machining deformation of biaxial residual stress.To address deformation control, Alvise et al.86employed the post-processing method of the tool path with deformation compensation.Jiang et al.87used the finite element method to predict the elastic deformation error in the machining process and proposed a novel scheme for modifying the machining path to compensate for the deformation error.They also analyzed the distribution of residual stress and the machining path on the machining surface and realized the control of cutting distortion and deformation of thin-walled components.Wellknown foreign enterprises, such as Starrag in Switzerland as well as Liechti and BCT in Germany, proposed an adaptive machining scheme based on online measurements.88–90In this scheme, the tool path compensation is based on the measured model, which demonstrates a certain deformation suppression effect during the machining of thin-walled components.

3.2.Dynamic response and chatter suppression of the timevarying machining system

At present, the research on machining chatter mainly includes chatter prediction and chatter suppression, as reported by Quintana and Ciurana.91In stability prediction,the dynamic modelling and parameter identification of the cutting system for complex thin-walled components provide important foundations for analyzing machining chatter.92–95Zhan et al.96established a three-axis machining dynamic model for the milling system and realized cutting stability boundary prediction.Schmitz and Donalson97proposed a response coupling substructure analysis method for frequency response prediction to realize a fast frequency response prediction of different tool types.Jin et al.98introduced an extended high-order time-domain method and established a prediction model by using the corresponding 3D stability lobe diagram.In addition, Postel et al.99predicted the frequency response of tools at different rotational speeds using the decoupling method and the Timoshenko beam element.Feng et al.100established a relationship between dynamic parameters and milling positions by combining their proposed method with the Kriging agent model and finite element simulation.Li et al.101analyzed the stability of the machining process and established a stable parameter domain, which were important ways of suppressing machining chatters.Altintas et al.102established a classical three-axis machining stability prediction model and a zeroorder frequency domain method, which provide a foundation for exploring milling stability.Wan et al.103considered the multi-mode of the machining system and then proved that multi-mode stability can be achieved by the boundary envelope of the single-mode stability under certain conditions.They also proposed a method for predicting the lobe diagram during the multi-mode milling stability state.

In chatter suppression strategies, numerous researches applied active or passive control strategies to realize the chatter control by altering dynamic characteristics of milling systems,contributing to the increase in area of stability during cutting processes.The occurrence of chatter could be then suppressed to improve the machining accuracy and efficiency, in terms of passive and active controls.Here, the common passive damping devices installed on machine tools include friction dampers,tuned mass dampers, etc.The working principle is shown in Fig.11.104The passive control mainly applies the additional equipment or increases the damping to absorb the vibration energy of the milling system, aiming at realizing the chatter suppression.Kim et al.105introduced a multi-fingered cylindrical friction damper inside the milling cutter to suppress the occurrence of milling chatter.The multi-fingered cylindrical damper and the milling cutter extrude each other under the action of centripetal force during milling process, dissipating vibration energy through the mutual friction.Nakano et al.106employed dynamic absorbers attached to a collet chuck, aiming at suppressing chatters.Chowdhury et al.107introduced the negative stiffness inertial amplifier tuned mass dampers,and proposed the exact closed-form expressions to optimize system parameters for the dampers using H2and H∞optimization techniques.Ma et al.108presented a two-DOF tuned mass damper installed on the shank of the micro-mill and rotates with the tool, and then optimized the turned mass damper parameters based on the dynamics modeling of the milling structure.

For active controls, the feedback control principle is applied to compensate the control actively from the external supply of energy.109–111As a dynamic system, the active control system consists of sensor module, control module and power amplifier module and the associated schematic diagram of an active control system is illustrated in Fig.12.Wang et al.112realized an adaptive control of machining vibration through the optimizations of clamping and cutting parameters.Here, the real-time damping or stiffness adjustment based on the multi-order modal stability model could be achieved after considering process damping and variable dynamic characteristics.They also realized vibration control during the machining process by adding actuators and sensors around the spindle and tools as reported by the Sandia National Laboratories.Moreover, a milling chatter suppression method based on adaptive vibration modes was proposed, wherein the milling vibration frequency could be modified and controlled in the frequency domain.113Brecher et al.114suppressed the production of machining chatter by installing an active fixture system on the CNC milling machine tool.On the other hand,the spindle speed variation(SSV)as one of the effective method to suppress regenerative chatter, is similar to variable pitch and variable helix angle.Here, the chatter suppression is achieved through the regeneration effect during the disturbance machining process.Nam et al.115described a novel SSV, maintaining a constant absolute acceleration rate to overcome the limitation of the conventional SSV.The chatter stability could be improved in all speed regions by keeping the absolute acceleration rata as a constant.Yamato et al.116represented a method of sinusoidal SSV,that is,SSSV.They expressed the analytical kinematic energy model with the Bessel function to dissipate the chatter energy effectively.Otto and Randons117applied a simplified theory for the stability analysis with slowly timevarying spindle speed to estimate the efficiency using SSV methods.

3.3.Adaptive machining and error control of complex thinwalled components

Adaptive machining as a comprehensive manufacturing technology integrates the functions of state perception, real-time analysis, independent decision-making, and accurate execution.The state of machining process can be determined by comparing the measurement data, and then the machining strategy is autonomously formulated according to the results,and finally the processing program suitable for the current machining state is generated.Moreover, the NC machining scheme can be customized for each personalized machining object by adjusting the processing program once the shape and position of the machining object change.For complex thin-walled components of aero-engine, the deformation prediction and error compensation should be also integrated as the adaptive machining method is applied.Prior to the machining process, the current status of blanks is obtained by means of measurement and the following NC processing parameters are adjusted by comparing the blank design model.Subsequently, the adaptive machining technologies are employed for the further digital measurements,adaptive workpiece clamping, registration and positioning, process model construction and NC machining programming.118–120After applying the abovementioned technologies, a timely adjustment can be achieved according to the current part deformation, uneven part allowance and inaccurate clamping position, aiming at adapting the current part state and completing the specific processing.General Electric in America realized an adaptive machining of complex castings by monitoring the machining process in real time and modifying the machining programme.Meanwhile,the Northwest Polytechnic University in China realized a self-adaptive repair of blisks and a self-adaptive processing of precision forged and rolled blades.121,122Gao et al.123at Nottingham University developed an adaptive solution to reconstruct the nominal geometry of complex aerospace components.An adaptive repair solution integrated with additive manufacturing and machining process was obtained on basis of the reverse engineering system.Chen et al.124from Dalian University of Technology has done many researches on adaptive machining and proposed an adaptive federate planning method to improve contouring accuracy with constraints of chord error and drive constraints.Li et al.125from Guangdong University of Technology has been done many researches on error compensation,and realized the adaptive and real-time compensation of the non-uniform surface deformation errors.In addition, the adaptive machining process was also reported using the dimension online inspection method,126,127distortion control and error compensation methods,128etc.

Fig.12 Schematic diagram of an active control system.

Multiple sources and manufacturing processes are widely adopted in machining complex thin-walled components,whose final quality can be directly improved by controlling the errors amongst different processes.129,130Error compensation based on anti-deformation is commonly applied to control the elastic deformation of complex thin-walled weak rigid structures.Bera et al.131predicted the machining error caused by tool and workpiece deflection based on the cutting force model with geometric changes and then compensated for the error by modifying the tool path.To control the error flow in multiple processes, Stryczek132proposed an error compensation method in the multidimensional decision space based on the vector equation.Li et al.133established a force-induced deformation prediction model based on the static substructure method and presented a flexible error compensation strategy for the deformation of thin-walled components and tools in the milling flank structure.Zuo et al.134combined an error compensation methodology for thin-walled components with the inverse reconstruction model, which integrates the compensation behaviour of workpiece into error suppression.Hou et al.135constructed a general model for the machining error compensation of thin-walled components and built a discrete control system for error compensation.They realized an offline learning of the compensation coefficient and an accurate control of the machining error by using the feedback principle.Bravo et al.136allowed on the possibility of both subsystems vibrating at the same time,and the method to predict this situation.Fig.13 is they proposed model of a single degree of freedom cutting process.Similarly, the model for milling is proposed, as shown in Fig.14.

Fig.13 Schematic representation of a single degree of freedom cutting process136.

Fig.14 Schematic representation of milling process136.

Numerous studies have also focused on controlling the machining accuracy of thin-walled components.However,these studies have mainly considered a single process and target.In this case,the machining accuracy control should be further studied under the influence of combined process coupling in combination with the complex profile and thin-walled structure characteristics of blisks and casings.To ensure the machining accuracy and efficiency of complex thin-walled components, the available methods should be combined with the deformation mechanism,chatter suppression and adaptive machining process.Furthermore,the machining accuracy control in the whole manufacturing process of complex thinwalled components warrants further investigation.

4.Machining surface integrity control technology

Apart from shape accuracy, the machining surface state (i.e.surface integrity) of complex thin-walled components in aero-engines is another core factor that affects their service performance in actual operations.Problems in the manufacturing quality of key components can be ascribed to the lack of a clear mechanism of surface state creation, insufficient research on the relationship between manufacturing surface state and service performance, lack of relevant basic data and failure to master the process control rules.In recent years,numerous studies have started exploring the characterization,generation, evolution and control of machined surface states137–139.

4.1.Characterization model and evaluation method of component surface state

Scientific and reasonable surface state characterization and evaluation serve as the mathematical and theoretical bases of subsequent research on surface state generation mechanism and process control.The characterization and evaluation of surface state are rooted in the concept of surface integrity.The Air Force Materials Laboratory in America has conducted investigations of machined surface integrity since 1948.After Field and Kahles proposed the concept of surface integrity in 1964, the national standards for surface integrity were proclaimed, and the associated process specifications and acceptance criteria for the key components were then formulated.140,141In characterizing manufacturing surface state,the machined surface state has been usually characterized by the stress concentration coefficients (SCF), surface residual stress, surface micro topography and surface threedimensional topography,etc.Cheng et al.142adopted the fractal function to reconstruct the machined surface profile and evaluated the stress concentration degree of the machined surface using stress concentration coefficients.They also applied the polynomial function143and sinusoidal decay function144to determine the gradient distribution of surface micromechanical characteristics using the parametric characterization method.Imran et al.145divided the cutting surface layer into the nanostructure surface layer, deformed secondary surface layer and material matrix and then evaluated the cutting surface layer using grain size, nanohardness, plastic deformation and crystal orientation differences.In surface state evaluation,the surface integrity dataset was introduced in America in 1972 as the standard for evaluating the mechanical surface state.Meanwhile, in China, researchers from the Northwestern Polytechnical University revealed that the surface stress concentration coefficient, surface material stress concentration sensitivity index and component fatigue strength coefficient can be used to evaluate the surface state.146,147Some typical methods for characterizing the machined surface state are illustrated in Fig.15.

4.2.Effect of component surface state on fatigue performance and service evolution

In recent years, numerous researchers have explored the influence of surface state on fatigue performance and service evolution due to the close relationship between the fatigue performance of components and the associated manufacturing surface state.148–151Peng et al.152found that the temperature and cyclic loading time severely influence the stress relaxation limit and rate of titanium alloys during the uniaxial tensile test(Fig.16).Meguid and Maricic153reported that the residual stress of Inconel 718 superalloys caused by abrasive blasting processes can be completely eliminated under the overloaded thermal force load within the first cycle.Cui et al.154studied the low cycle fatigue performance of M951G alloys under different total strain amplitudes at 900℃and 1000℃and found that the precipitation phase of γ’is related to the testing temperature and strain amplitude, thereby affecting the fatigue performance.The deformation mechanism for the M951G alloy is illustrated in Fig.17.

In terms of the influence of surface state on fatigue performance, Zhang et al.155characterized the variations in the microstructures of titanium alloys before and after abrasive blasting processes and found that the grain refinement and increase in grain boundaries contribute to the prevention of microcrack propagation and improvement in fatigue life.Shi et al.156found that the fatigue performance of TC17 titanium alloys can be significantly improved by using vibrating polishing technology in the conventional abrasive blasting process.Yang et al.157observed a remarkable reduction in surface roughness after the ultrasonic rolling of GH4169 superalloys,which generates residual compressive stress and grain refinement in the surface layer, thereby greatly improving the fretting fatigue life.

Fig.15 Typical characterisation methods for the machined surface state.

Fig.16 SR curves of Ti–6Al–4 V titanium alloy samples under different conditions152.

Fig.17 Schematic illustrations of the deformation mechanism of the M951G alloy154.

4.3.Manufacturing surface state generating mechanism and process control method

Fig.18 Grain refinement mechanism of nickel-based superalloys by severe plastic deformation160.

The generating mechanism of the surface state refers to the influence mechanism,model and rule of a single process formation of a surface state(e.g.electrolysis,cutting,grinding and strengthening) and the composite reconstruction of multiple processes(e.g.electrolysis-cutting-grinding-strengthening).158,159In terms of the formation mechanism and influence rule of cutting surface states,Liao et al.160revealed the grain refinement mechanism of nickel-based superalloys under high strain rate and temperature(Fig.18).Kalisz et al.161analysed the surface state characteristics after a continuous milling/polishing of curved surfaces based on new technologies and tribological principles.Thakur and Gangopadhyay162studied the influence of cutting parameters and tool wear on the surface microstructure of superalloys.In terms of the surface state formation and process control of abrasive blasting process, Byrne et al.163comparatively studied the surface of titanium alloys treated with abrasive blasting processes and found that the surface roughness increases by 75 % after the abrasive blasting process.Maleki et al.164studied the effects of three strengthening processes(i.e.high-intensity shot peening, laser impact strengthening and ultrasonic nanocrystal surface modification) on the microstructure,mechanical properties and fatigue behaviour of Inconel718 superalloys.The cooling conditions also significantly affected the formation of surface integrity of the machined surface.Current research hotspots for enhancing the surface quality of difficult-to-machine materials preferred the use of nano-enhanced biolubricants (NEBL) and electrostatic atomization minimum quantity lubrication (EMQL).The polar groups of biolubricants were conducive to the formation of biofilms in the cutting area,and the higher viscosity and surface tension were also conducive to lubrication.165,166The excellent anti-friction and anti-wear properties of nanoparticles could improve the tribological properties of the cutting interface.In addition, the nanoparticles have filling effect on the machined surface.167At present, NEBL has been widely used in turning,milling and grinding of difficult-to-machine materials.168,169Extensive experiments also confirmed the improvement of surface roughness by NEBL.Scientists exploring the action patterns of NEBL found that mixing nanoparticles and mixing vegetable oils seemed to improve surface quality more.170Zhang et al.171explored the use of hybrid nanofluids to improve the surface quality of workpieces from difficult-tocut materials in MQL grinding.The results show that the use of 6wt% MoS2/CNTs (mixing ratio 2:1) mixed nanoparticles can generate Raof 0.311 μm.The technology could be extended to a variety of machining processes for difficult-to-machine aerospace materials to address the bottleneck of insufficient surface integrity.Compared to pneumatic atomization MQL,EMQL was able to reduce the surface roughness of titanium alloys due to its unique empowering mechanism and atomization properties.In addition, the scientists also verified that lecithin could be used as an additive to improve the atomization and machining performance of electrostatic atomization.172To solve the problem of thermal damage to workpieces under extreme machining conditions of difficult-to-machine materials,scientists have performed a combination of cryogenic air and MQL.173This multi-process coupled lubrication seemed to have more potential for development.

Many studies on the characterisation and evaluation of surface states,anti-fatigue mechanism and evolution,surface generating mechanism and process control of complex thin-walled components have merely focused on a single process.In these studies, certain problems, such as the superposition, coupling and reconstruction of surface states under integrated multiple processes as well as the complex structure characteristics of thin-walled components, should be systematically considered.The innovative combining processes for fabricating blisks involve the low-stress and high-efficiency electrolytic premachining combined with precision cutting, precision polishing and surface strengthening.Meanwhile,manufacturing processes with high-efficiency rough machining combined with ultrasonic finish machining can be used in machining thinwalled casings.Therefore, the anti-fatigue mechanism of the surface state should be further investigated using the digital characterization method, and the generating and refactoring manufacturing mechanisms should be applied to complex thin-walled components, such as blisks and thin-walled casings.A manufacturing process control method that combines multi-processes can be subsequently established to realize process inversion and parameter control from fatigue performance to surface state and to manufacturing.

5.Intelligent process control technology

The manufacturing accuracy and surface state of components are controlled by the coupling effects of multiple factors.However, the internal relationship between the process and the structure shape/surface integrity state under time-varying working conditions cannot be easily reflected.Therefore, the process integrative control of structure shape and surface integrity cannot be easily realized.Digital twin has been recently identified as a necessary means for the traditional manufacturing industry to transform itself from digital to intelligent.174–181This integration is also inevitable for physical entities to realize digital image and intelligent regulation.In addition, artificial intelligence is the critical technology during constructing digital twin model for machining cells.

In 2003, Michael Grieves from the University of Michigan proposed the concept of digital twin, which creates a virtual model of physical entities in a digital way.182–186Subsequently,the interaction and integration in physical and information worlds can be achieved with the help of data to simulate the behaviour of physical entities using several methods,including virtual real interaction feedback,data fusion analysis and decision iteration optimization.The service life of aircraft construction can also be predicted by combining digital twin with the aircraft virtual model and the structural deviation and temperature calculation model, which can affect flight parameters according to the Air Force Research Laboratory in 2011.187Since then,digital twin has received much academic and industrial interest at home and abroad.188,189However,the development and application of digital twin remain at their infancy.The existing technologies need to be explored further,and high-fidelity simulation methods should be developed to achieve a full-scale intelligent optimization and control of complex systems.

In recent years, the virtual twin and intelligent control of processing technology have received many attentions by combining the traditional processing technology with artificial intelligence.190–192Zhao et al.193proposed the dynamic optimization method of cutting parameters by exploring the intelligent reconfiguration of the cutting process and the dynamic optimization method of cutting parameters.Applying this model could ensure a smoother cutting force of the optimized cutting process, avoid the excessive deformation and unexpected vibration of machine tools and guarantee excellent equipment performance.The digital twin principle of the cutting process is illustrated in Fig.19.

Li et al.194proposed an online monitoring method for tool wear in numerical control milling driven by digital twin.A real-time collection of dynamic data is realized to monitor tool wear status, simulate the tool wear process and formulate a strategy of tool changes.This method reduces the amount of time and resources consumed in model training, thereby satisfying the real-time requirements in the machining process.However, despite efficiently predicting tool wear, this method is unable to provide a real-time feedback and timely control of the physical machining process.Cao et al.195built a new machining simulation system to address the problem where the simulation system refuses to cooperate with the physical machining process and then developed the corresponding optimization algorithm.Oyekan et al.196developed an automatic unit based on visual sensing technology and digital twin technology to solve the problem of blade scrapping caused by workers repairing the blades and then established the digital twin model of blade grinding parameters.This automatic unit continuously tracks and removes the coating material from the surface of the fan blade in a closed-loop manner and addresses the blade wear problem, thereby improving the maintenance efficiency and reducing the maintenance costs.However, they did not consider embedding the knowledge of skilled workers into the automation unit.Xie et al.197proposed a cutting tool digital twin model for the all-life cycle based on the interaction and fusion between physical tool entity and its virtual counterpart,hence guiding the future development of intelligent manufacturing.Liu et al.198proposed a digital twin modelling method based on the bionic principle and developed multiple interactive digital twin models that illustrate the complete physical machining process.

The above research provides a solid theoretical basis for the intelligent control of traditional processing technologies.However, the cutting, polishing and strengthening manufacturing processes of complex thin-walled components in aero-engines are nonlinear multi-field coupling machining processes.199–201Along with material removal and surface formation, a variety of complex physical processes(e.g.deformation and vibration)are highly coupled, thus affecting the accuracy and surface state of the target component.202–204China is yet to achieve an intelligent process control of complex thin-walled components.Therefore, a twin model of the whole manufacturing process of complex thin-walled components should be built to realize structure shape and surface integration adaptive control under the complex component combination manufacturing process.205–209Wang et al.210provided the architecture of digital twin for digital manufacturing that can support the operation and manufacturing service for optimizing operations and failure prediction (Figs.20 and 21).With the continuous updating of digital twin driven by real data, the machining state prediction and control of machining accuracy and surface state for complex thin-walled components can be further realized, which can significantly improve the machining efficiency and quality of key components of aero-engines and ensure the safety of the machining process.

In digital twin model,artificial intelligence is a critical technology to improve the manufacturing performance of systems,aiming at realizing the autonomous perception, autonomous decision-making and accurating implementation of processing systems in twin models.However, the application of digital twin in process control is still at the exploratory stage.Specifically, the twin data are insufficient and inaccurate, and then the integration of physical entity data, virtual model data and service application data is difficult to realize.211In addition, the quality prediction and control of complex constructions based on simulation method relies on ‘shape similarity’than on ‘spirit similarity’, failing to meet the requirements of the bidirectional mapping of twin models in digital twin.In this case, the real-time information interaction between physical entities and digital twins also needs further improvements212,213.

Fig.20 Architecture of digital twin for digital manufacturing210.

Fig.21 Digital-twin-enabled fault diagnosis framework210.

At present,the control of relevant physical quantities is also a research hotspot,in terms of chatter detection,tool condition monitoring,and even multi-condition identification,which can be employed to manufacture complex thin-walled components.125,214Meanwhile, the combination of machine learning and neural network can realize the intelligent process control based on the abovementioned online detection technology.Rahimi et al.215proposed a hybrid chatter detection model based on artificial neural-network learning system.They separated the forced vibrations and chatter,and detected the occurrence of chatter and its frequency.Schueller and Salda?a216conducted milling tool life experiments under various machining conditions, and collected the sound, spindle power and axial load signals to prediction the tool wear behavior using a machine learning model.Karandikar et al.217applied Bayesian machine learning and performed experiments to describe automated identification of the milling stability boundary.The adaptive experimental strategy was presented to identify the stability boundary.

6.Status, challenge and tendency

At present, the high degree of coupling with multiple process parameters and the associated factors (e.g.force, heat and field) has a reciprocal effect on material removal behaviour in the manufacturing of complex thin-walled components of aero-engines.In this case, researchers are unable to describe the coupling relations between various variables and the machining accuracy and surface integrity of complex thinwall components due to their extremely complex timevarying and dynamic behaviour in the processes of surface formation and surface generation.In response to this restriction,manufacturing accuracy or surface integrity controls are usually considered independently in previous studies.Therefore,the development of integrated structure shape and surface integrity control manufacturing processes cannot be accommodated due to the lack of intelligent process models and means for controlling machining accuracy and surface integrity collaborative manufacturing.The deep analysis of the development status and tendency of related research can highlight some significant advancements in the manufacturing technology of key components of aero-engines.Fig.22 illustrates the status and development tendency on collaborative manufacturing technologies of structure shape and surface integrity for complex thin-walled components, in terms of material removal mechanism, precision control of structure shape,machined surface integrity control and intelligent process control technology.

In material removal mechanism, at present, previous researches have mainly focused on simple structures under single energy field (e.g.traditional cutting, electrolysis and ultrasonic cutting, etc.).Here, the material removal and surface formation mechanisms cannot be fully revealed,owing to lacking the comprehensive influences of complex profile structures on certain variables (e.g.force, heat, and flow, etc.) under the multiple energy coupling field.In this case,the challenges have to be faced to obtain the desired machining efficiency and quality requirements for complex thin-walled components of aero-engines.Therefore, in the future, the strong coupling effects (e.g.force, heat, etc.) should be studied after considering the complex structures of aero-engine components under multiple energy fields.In addition, the mapping model amongst machining parameters,forming efficiency and surface quality can be established.Subsequently,the integrated manufacturing design processes will be created on basis of the research findings of the material removal mechanism, providing an advanced process method and process flow support for the subsequent improvements in machining accuracy control, surface integrity control and collaborative intelligent manufacturing.

In machining accuracy control, the current researches put the emphasis on multi-factor coupling and process signal sensing processes, whereas the multiple data coupling investigations are unable to conduct due to the analyzing the data of a certain factor.Therefore, the applications on machining accuracy control technology are severely affected on basis of the current research points.In the future, on one hand, the analysis of collected coupling signals should be enhanced by developing multi signal fusion methods, and then the data after fusion analysis and processing can be used as the input of the intelligent algorithm.Subsequently,the influence degree of each factor can be finally output through the intelligent algorithm during machining processes,and then the machining accuracy control scheme can be independently determined.On the other hand, the evolution rules of integrated manufacturing processes related to structure formation and surface integrity should be evaluated,and the associated stability prediction and chatter control strategy should be developed using the established dynamic modelling and parameter identification method.The adaptive manufacturing accuracy control of complex thin-walled components can be realized based on the proposed deformation mechanism of processing elastoplasticity and residual stress.

In surface integrity control, at present, the increasing complexity of aero-engine components introduces further challenges in materials processing.The combined machining process has gradually developed from traditional manufacturing methods relying on a single process, aiming at improving machined surface quality and efficiency, as well as reducing costs.For example, the integrated manufacturing process involving electrolytic high-efficiency rough machining, milling finishing, precision grinding and polishing and surface strengthening can be adopted for machining blisks.Previous studies on manufacturing accuracy and surface state have mainly focused on a single process without considering the reconstruction rules and mechanisms.Moreover, the characterization of machined surface state is mainly directed at a single point, which is difficult to describe the machined surface state comprehensively.Obviously, the urgent demands for the manufacturing development of complex thin-walled components fail to satisfy.Therefore, the multi-dimensional characterization and formation mechanism of the machined surface should be further developed to realize the machined surface integrity control.Meanwhile, the digital characterization modelling and evaluation methodology of surface integrity should be established along with the mechanisms of surface state generation and reconstruction of composite antifatigue surface state under the force–heat–structure coupling effect in the machining process.A manufacturing process control method based on the anti-fatigue surface state should also be established to provide a theoretical guarantee for the formulation and implementation of a collaborative control process strategy for complex thin-walled components.

Fig.22 Status and development tendency on collaborative manufacturing technologies of structure shape and surface integrity for complex thin-walled components.

In intelligent process control, the machining control methods can be obtained through various proposed intelligent algorithms.However, many steps of the machining process and implement are still participated by human,introducing further challenges to the intelligence of intelligent algorithms.In this case, the ability of autonomous decision-making and accurate execution of intelligent algorithms should be strengthened.In the future, the theory and method of digital twin accurate modelling in multi-dimensional and multi-scale manufacturing processes need to be studied based on the manufacturing accuracy and surface integrity control theory of complex thin-wall components.The physical space, information space, accuracy and surface integrity can be combined to form a multidimensional twin agent that realizes a real-time synchronous simulation and virtual real linkage control of multiple elements, objectives and services in the manufacturing space.Industrial big data analysis and applications oriented to the time-varying working conditions, structure shape and surface integrity should also be studied to respond to certain problems,such as the variety of types, high complexity, large amount of process data and low utilization rate of massive data in machining complex thin-walled components.Intelligent decisions related to machining accuracy and surface integrity coordination can be made based on deep learning technology.

7.Conclusions

(1) The collaborative manufacturing technology of structure shape and surface integrity exhibits numerous advantages as a processing method, such as high flexibility, efficiency and precision and low cost.This technology has broad application prospects in machining difficult-to-cut materials of complex thin-walled components in aerospace and provides theoretical and technical supports for the manufacturing of complex surface components with high reliability.

(2)Benefits in quality,efficiency and cost can be simultaneously realized by adopting the integrated machining processes,which involves cutting, electrolysis, grinding, polishing and ultrasonic cutting.The mapping and inversion models amongst the machining parameters,energy field characteristics and material removal/surface generation should be established to provide a theoretical support for the formulation of a machining accuracy and surface integrity control strategy for complex profile components.

(3)The generation mechanism of the complex blank state of components is revealed based on machining dynamics and deformation prediction models.The machining process can also be analyzed in real time and autonomously decided by sensing the real-time changes in the machining parameters with multi-source sensors.The machining precision control mechanism of the complex blank state of components can then be revealed, hence providing a basic model and process method for realizing the machining accuracy control of complex thin-walled components.

(4) The reconstruction mechanism of machined surface states under multiple processes is revealed by combining the manufacturing process factors, force thermal structure coupling effect, stress–strain effect and manufacturing surface state system.The regulatory mechanism of machined surface states is then studied after analyzing the control ability,surface state laws of manufacturing process factors and the evolution law of the surface state,hence providing a theoretical basis for developing and establishing the control technology of complex thin-walled components.

(5) An across-scale digital twin models based on the whole manufacturing process can be constructed by combining the manufacturing process parameters, stress distribution state,force–heat–structure coupling effect and physical mapping relationship between the material properties of the workpiece and the shape properties of the target components.A quantitative is then established relationship amongst the process parameters,processing mechanism and instantaneous shape of components to provide the model with some basis and intelligent algorithm support to realize a collaborative control of the structure shape and surface integrity of complex thin-walled components from the perspectives of macroscopic manufacturing, mesoscopic morphology and microscopic evolution.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgments

This work was financially supported by the National Natural Science Foundation of China (Nos.51921003, 92160301,52175415 and 52205475), the Science Center for Gas Turbine Project(No.P2022-A-IV-002-001)and Natural Science Foundation of Jiangsu Province (No.BK20210295).