Robotic Grinding Process Research Articles

Measurements of Geometrical Quantities and Selection of Parameters in the Robotic Grinding Process of an Aircraft Engine

Measurements of Geometrical Quantities and Selection of Parameters in the Robotic Grinding Process of an Aircraft Engine

Process Optimization of Robotic Grinding to Guarantee Material Removal Accuracy and Surface Quality Simultaneously

Abstract Simultaneously guaranteeing material removal accuracy and surface quality of robotic grinding is crucial. However, existing studies of robotic grinding process optimization have mainly focused on a single indicator that solely considers contour error or surface roughness, while studies that simultaneously investigate the impact of contact force, spindle speed, feed rate, inclination angle, and path space on the material removal profile (MRP) and the surface roughness are lacking. This paper proposes a hybrid optimization method that considers dimensional accuracy and surface quality constraints. First, an MRP model that considers the coupling influence of the contact force, spindle speed, feed rate, and inclination angle is presented. Then, a surface roughness model that considers the inclination angle is established. Finally, the contact force, feed rate, inclination angle, and path space are simultaneously optimized to satisfy the hybrid constraints of MRP accuracy and surface roughness. The proposed method ensures maximum grinding efficiency while satisfying dimensional accuracy and surface quality constraints. The proposed method is verified on an industrial robotics grinding system with a pneumatic force-controlled actuator. The results show that the proposed method has higher profile accuracy and lower surface roughness than traditional methods.

Robotic Grinding Process of Turboprop Engine Compressor Blades with Active Selection of Contact Force

Robotic Grinding Process of Turboprop Engine Compressor Blades with Active Selection of Contact Force

Robotic Grinding Process Monitoring by Vibration Signal Based on LSTM Method

In robotic grinding, the trajectory deviation of the robot will change the grinding force and then affect the grinding material removal. Therefore, grinding force and material removal monitoring is important for the machining quality of robotic grinding. In this paper, force and material removal estimation methods are proposed based on the acceleration signal. The dynamic model of robotic grinding is studied and the vibration force relationship can be obtained. Based on the vibration force relationship, the framework of the grinding force and material removal monitoring is established. In this framework, the peak value of the acceleration is firstly extracted by the signal process, and the Long Short-Term Memory (LSTM) network is used to train the vibration-force mapping model. Then the grinding removal can be calculated by the empirical formula. The experiments of robotic grinding are carried out. The real grinding force and the material removal can be estimated in time, which proved the effectiveness and potential of our method.

Investigation of cycle time behavior in the robotic grinding process

The production of power transmission components such as gears and bearing rings involves precision finishing of highly specialized and dedicated precision grinding machine tools. Six degree-of-freedom (6-dof) articulated arm industrial robots are less expensive and more versatile than precision grinding machines because they can be readily reconfigured for a variety of production tasks including precision grinding of free-form surfaces. However, the high compliance of such robots poses a challenge in high-precision operations, particularly as it relates to the overall cycle time of the process. The paper presents an analysis of the effects of robot compliance and grinding parameters on the cycle time of a precision grinding process. Specifically, a process model for robot-workpiece interaction and grinding process cycle time is developed. This model is used to analyze the behavior of a face grinding process cycle. The results show that for a given robot stiffness, minimizing the face grinding cycle time requires the process to be operated at the highest achievable wheel speed and infeed rate with a wheel of higher hardness. The paper also shows that a 6-dof robot offers the flexibility of modifying the Cartesian stiffness at the end effector by varying the robot pose, which is difficult to achieve in conventional grinding machines. This capability, along with the process cycle time model, permits optimization of process cycle times for precision grinding of surfaces, including free-form surfaces, using 6-dof robots.

A novel material removal prediction method based on acoustic sensing and ensemble XGBoost learning algorithm for robotic belt grinding of Inconel 718

High-performance component manufacturing has increasing needs of robotic grinding process that can achieve accurate material removal. This article proposes a novel material removal model for robotic belt grinding of Inconel 718 based on acoustic sensing and machine learning. The sound signal is collected online by an audio sensor during the grinding process. A novel method to identify the idle running period and eliminate noise is developed using discrete wavelet decomposition (DWD) and fast Fourier transformation (FFT). Statistical features are extracted from each clean acoustic signal segment to better represent and quantify grinding process. A new k-fold eXtreme Gradient Boosting (k-fold-XGBoost) algorithm after training and optimization is integrated into the material removal (MR) model. The test results indicate that the values forecasted by the model are consistent with the measured values. The mean absolute percentage error (MAPE) of material removal evaluated by the model is 4.373%, which shows a better performance than the reported results which are in the range of 6.4 to 8.72%. In comparison with other prediction models, such as optimally pruned extreme learning machine and random forest and support vector regression, k-fold-XGBoost model shows superior results for the same datasets. It can be concluded that the proposed method based on acoustic signal and the ensemble learning model is effective in predicting the material removal despite the complicated grinding environment.

Application of novel force control strategies to enhance robotic abrasive belt grinding quality of aero-engine blades

Robotic belt grinding has emerged as a finishing process in recent years for machining components with high surface finish and flexibility. The surface machining consistency, however, is difficult to be guaranteed in such a process. To overcome this problem, a method of hybrid force-position control combined with PI/PD control is proposed to be applied in robotic abrasive belt grinding of complex geometries. Voltage signals are firstly obtained and transformed to force information with signal conditioning methods. Secondly, zero drift and gravity compensation algorithms are presented to calibrate the F/T transducer which is installed on the robot end-effector. Next, a force control strategy combining hybrid force-position control with PI/PD control is introduced to be employed in robotic abrasive belt grinding operations where the force control law is applied to the Z direction of the tool frame and the positon control law is used in the X direction of the tool frame. Then, the accuracy of the F/T transducer and the robotic force control system is analyzed to ensure the stability and reliability of force control in the robotic grinding process. Finally, two typical cases on robotic belt grinding of a test workpiece and an aero-engine blade are conducted to validate the practicality and effectiveness of the force control technology proposed.

Model Predictive Control of Robotic Grinding Based on Deep Belief Network

Considering the influence of rigid‐flexible dynamics on robotic grinding process, a model predictive control approach based on deep belief network (DBN) is proposed to control robotic grinding deformation. The rigid‐flexible coupling dynamics of robotic grinding is first established, on the basis of which a robotic grinding prediction model is constructed to predict the change of robotic grinding status and perform feed‐forward control. A rolling optimization formula derived from the energy function is also established to optimize control output in real time and perform feedback control. As the accurately model parameters are hard to obtain, a deep belief network is constructed to obtain the parameters of robotic grinding predictive model. Simulation and experimental results indicate that the proposed model predictive control approach can predict abrupt change of robotic grinding status caused by deformation and perform a feed‐forward and feedback based combination control, reducing control overflow and system oscillation caused by inaccurate feedback control.

叶片复杂曲面的机器人抛磨工艺规划

叶片复杂曲面的机器人抛磨工艺规划

Variable Speed Control Using Fuzzy-PID Controller for Two-phase Hybrid Stepping Motor in Robotic Grinding

In a robotic grinding process, a light-weight grinder is held by an intelligible robot arm. Material removal is carried out by the rotating grinding tool while the end effector of robot guarantees that the tool follows a programmed path in order to work on complex curved surfaces. Grinding tool is driven by Two-Phase Hybrid Stepping motor derive. This work aims to develop a controller based on fuzzy logic to improve the speed control performance of Two-Phase Hybrid Stepping motor derive in order to achieve a controller that provides grinding with higher quality. The analysis and design of PID-Fuzzy controller to improve the response of the motor speed were studied. This paper simulates six motor speed input conditions. The simullink package of the MATLAB. Comparison between the conventional PID controller and Fuzzy-PID output was done on the basis of the simulation result obtained by MATLAB. The simulation results demonstrate that the designed Fuzzy-PID controller realize a good dynamic behavior of the Two-Phase Hybrid Stepping motor, a perfect speed tracking with less rise and settling time, minimum overshoot, minimum steady state error and give better performance compared to conventional PID controller.

Modal analysis of a light-weight robot with a rotating tool installed at the end effector

This paper investigates vibration of a moving flexible robot through modal analysis and by constructing vibration spectra of operational signals. A vector autoregressive model combined with a sliding window technique is used for signal processing in order to take into account system nonstationarity. Modal decomposition is conducted on the state matrix constructed from the appropriate vector autoregressive model parameters. A complete modal decomposition and spectrum construction algorithm able of highlighting the structural modes and harmonic excitations is presented. Through accurate identification from the vector autoregressive model, the method presented is able to discriminate, display and monitor the harmonics and structural modes during the processes investigated. This method is validated first by numerical simulation and then experimentally with a flexible robot performing three processes: moving a manipulator through the workspace, steady rotation of a grinder on the end effector and moving the manipulator combined with rotating the grinder. It is found on the operating robot that participation of the first structural mode is negligible when rotating the grinder but must be taken into account when moving the manipulator. The analysis presented and results obtained provide a sound basis for further investigation of vibroimpact behaviour in a robotic grinding process.

Force model for impact cutting grinding with a flexible robotic tool holder

Traditionally, grinding is used as a finishing process in the manufacturing chain. However, in recent years, it has been also used to machine hard or brittle materials. Another application, developed by Hydro-Quebec’s research institute, IREQ, is robotic grinding for in situ maintenance of large hydropower equipment. This high material removal rate grinding process is performed with a flexible robot as the tool holder. In a robotic grinding process, having an accurate model to predict the process force is crucial in order to achieve the desired material removal rate. This paper presents an experimental study that breaks new ground in determining the coefficients of an existing semi-analytical force model. An impact cutting behavior has been clearly observed at various grinding power levels both with high-speed video and force-measuring equipment. The force model is based on an idealized uncut chip at each wheel impact on the surface. The number of impacts per revolution of the grinding wheel is ascertained and used to determine the coefficients of the force model. The previously determined energy partition ratios and correlated model are also used to determine the ratio of energy consumption through friction to that for chip formation. The force coefficients are then fine-tuned and updated using the energy partition model and friction-chip energy ratio for the process. The newly determined coefficients are validated through a series of tests and shown to be in good agreement with measured grinding forces. The results show that the new enhancements in determining model parameters can be used to better predict the process force, power, and depth of cut in different cutting regimes and at various grinding power levels.

Characterization of grinding wheel grain topography under different robotic grinding conditions using confocal microscope

Knowing the grain geometry in grinding wheels is an asset for better understanding the grinding processes. This study investigates the grain protrusion and rake angles of two self-dressing zirconia-alumina grinding wheels in a robotic grinding process. The topography of the wheel is measured using a confocal scanning laser microscope. An optical image of the surface is used to create a mask of the grains with image processing techniques. Grain geometry information is then obtained by applying the mask to the entire surface. A vertex normal technique is used to find the cutting edges facing the cutting direction and only consider those edges in grain rake angle calculations. Surface parameters, including grain density, width, protrusion height, and rake angle, are extracted from the topography. The grinding wheel is characterized in low, medium, and high depths of cut in the range of robot operation. Results indicate that grain density, width, and protrusion height distribution are not affected by the depth of cut. It is also found that in shallow grinding, grain rake angle shifts slightly to higher negative angles; whereas, with a higher depth of cut, sharper edges exist on the wheel surface, which improve process efficiency.

Vibro-impact dynamics of material removal in a robotic grinding process

This paper studies the vibratory dynamics governing material removal performed by a compliant robot. The objective is to understand whether metal removal at a target rate or to a target depth of cut is possible given the unavoidable significant and sustained vibrations inherent to the process. Robotic grinding by the SCOMPI robot developed at Hydro-Quebec’s research institute for field maintenance work on hydroelectric equipment is studied. A test rig is developed to investigate the instantaneous process of material removal. The setup employs the methodology of angular analysis to detect and locate the discrete cutting events with respect to the instantaneous angular position of the spindle. Test results confirm that the grinding wheel exhibits cyclic impacting oscillations while removing material. This vibro-impact behavior is explained by the nonsmooth nature of the system arising from the substantial difference between process stiffness and stiffness of the robot. Grinding force and power are formulated based on the impacting dynamics of material removal (impact cutting). Parameters of the impact-cutting model are determined experimentally. The impact-cutting model is found well-suited to predicting the grinding power required to remove metal at a target rate. The conclusion is reached that metal removal to a target depth and with acceptable surface waviness is possible despite high-amplitude vibro-impacts between the grinding wheel and workpiece.

Experimental and finite element analysis of temperature and energy partition to the workpiece while grinding with a flexible robot

Grinding processes performed with flexible robotic tool holders are very unlike conventional types of grinding because of low stiffness of the robot's structure. A special flexible robotic grinding process is used for in situ maintenance of large hydroelectric equipment for bulk material removal over large areas rather than as a finishing step, as is the case for most conventional grindings. Due to the low structural stiffness of tool holder, cutting is interrupted at each revolution of wheel during the grinding process. In this study, an investigation is carried out to determine the temperatures and energy partition to the workpiece for the above-mentioned flexible robotic grinding process by a three-dimensional finite element thermal model. Experiments were undertaken using embedded thermocouples to obtain the subsurface temperature at several points in the workpiece during the process. Then, energy partition to the workpiece was evaluated using a temperature-matching method between the experimental and numerical results. This ratio is used for predicting the temperature field at the wheel–workpiece interface with a relevant heat source function. Kinematics of cut and the flexible robot's dynamic behavior are considered in applying the heat input to the model. The energy partition to the workpiece in this specific flexible grinding process is found to be lower than for analogous conventional precision grinding processes. Two models, one from the literature and one from the power model of the process, are modified and proposed for determining the energy partition. The results showed that the energy partition ratio decreases by increasing the process power. Also, this ratio slightly decreases at higher feed speeds. In addition, lower temperatures were seen at higher powers due to the lower intensity of heat input over a larger contact area. Experimental observations show close agreement between simulated contact temperatures and measured results.

Angular analysis of the cyclic impacting oscillations in a robotic grinding process

In a robotic machining process, a light-weight cutter or grinder is usually held by an articulated robot arm. Material removal is achieved by the rotating cutting tool while the robot end effector ensures that the tool follows a programmed trajectory in order to work on complex curved surfaces or to access hard-to-reach areas. One typical application of such process is maintenance and repair work on hydropower equipment. This paper presents an experimental study of the dynamic characteristics of material removal in robotic grinding, which is unlike conventional grinding due to the lower structural stiffness of the tool-holder robot. The objective of the study is to explore the cyclic nature of this mechanical operation to provide the basis for future development of better process control strategies. Grinding tasks that minimize the number of iterations to converge to the target surface can be better planned based on a good understanding and modeling of the cyclic material removal mechanism. A single degree of freedom dynamic analysis of the process suggests that material removal is performed through high-frequency impacts that mainly last for only a small fraction of the grinding disk rotation period. To detect these discrete cutting events in practice, a grinder is equipped with a rotary encoder. The encoder's signal is acquired through the angular sampling technique. A running cyclic synchronous average is applied to the speed signal to remove its non-cyclic events. The measured instantaneous rotational frequency clearly indicates the impacting nature of the process and captures the transient response excited by these cyclic impacts. The technique also locates the angular positions of cutting impacts in revolution cycles. It is thus possible to draw conclusions about the cyclic nature of dynamic changes in impact-cutting behavior when grinding with a flexible robot. The dynamics of the impacting regime and transient responses to impact-cutting excitations captured synchronously using the angular sampling technique provide feedback that can be used to regulate the material removal process. The experimental results also make it possible to correlate the energy required to remove a chip of metal through impacting with the measured drop in angular speed during grinding.

Study of Workpiece Temperature Distribution in the Contact Zone during Robotic Grinding Process Using Finite Element Analysis

Grinding is traditionally categorized as a finishing process in manufacturing. However, more recently it has been used as a machining process as well. Temperature distribution is crucial for investigation of thermal softening effects in the material removal process and thermal damages on the surface of workpiece. One important aspect to be considered in thermal simulations of robotic grinding is the dynamic behavior of the robot which can have significant influences on the process, especially in those performed with low stiffness robots. Most of the earlier thermal simulations of grinding processes are based on a simplified heat sourc e function representing the grinding wheel effect. In this study, heat generation due to a robotic grinding operation is distributed based on the local chip thickness and friction effect over the corresponding contact zone at the workpiece interface. The calculation of the chip thickness is based on a wear model of the grinding wheel in accordance with an impact-cutting behavior, observed with a high speed camera in our laboratory. Temperature distribution in the workpiece is simulated with a 3D transient thermal finite element (FE) code. Special attention is given to consider the dynamic effect of impact-cutting on the process which is caused by the high rotational speed of the wheel and low stiffness of the robot as a tool holder. Element deletion technique is used to represent the material removed from the workpiece and a well known model of the energy partition ratio is used and modified for the amount of energy entering into the workpiece. Grinding experiments conducted with a flexible robot showed a good agreement among simulation results and measured temperatures.

Designing and Optimization of an Off-line Programming System for Robotic Belt Grinding Process

Off-line programming (OLP) system becomes one of the most important programming modules for the robotic belt grinding process, however there lacks research on increasing the grinding dexterous space depending on the OLP system. A new type of grinding robot and a novel robotic belt grinding workcell are forwarded, and their features are briefly introduced. An open and object-oriented off-line programming system is developed for this robotic belt grinding system. The parameters of the trimmed surface are read from the initial graphics exchange specification (IGES) file of the CAD model of the workpiece. The deBoor-Cox basis function is used to sample the grinding target with local contact frame on the workpiece. The numerical formula of inverse kinematics is set up based on Newton's iterative procedure, to calculate the grinding robot configurations corresponding to the grinding targets. After the grinding path is obtained, the OLP system turns to be more effective than the teach-by-showing system. In order to improve the grinding workspace, an optimization algorithm for dynamic tool frame is proposed and performed on the special robotic belt grinding system. The initial tool frame and the interval of neighboring tool frames are defined as the preparation of the algorithm. An optimized tool local frame can be selected to grind the complex surface for a maximum dexterity index of the robot. Under the optimization algorithm, a simulation of grinding a vane is included and comparison of grinding workspace is done before and after the tool frame optimization. By the algorithm, the grinding workspace can be enlarged. Moreover the dynamic tool frame can be considered to add one degree-of-freedom to the grinding kinematical chain, which provides the theoretical support for the improvement of robotic dexterity for the complex surface grinding.

Automated robotic grinding by low-powered manipulator

A new robotic grinding process has been developed for a low-powered robot system using a spring balancer as a suspension system. To manipulate a robot-arm in the vertical plane, a large actuator torque is required due to the tool weight and enormous gravity effect. But the actuators of the robot system always exhibit a limited torque capacity. This paper presents a cheap and available system for precise grinding tasks by a low-powered robot system using a suspension system. For grinding operations, to achieve position and force-tracking simultaneously, this paper presents an algorithm of the hybrid position/force-tracking scheme with respect to the dynamic behavior of a spring balancer. Material Removal Rate (MRR) is developed for materials SS400 and SUS304. Simulations and experiments have been carried out to demonstrate the feasibility of the proposed system.

Parametric modelling and control of the robotic grinding process

The modelling and control of the grinding force for the robotic disc-grinding process are described. Modelling experiments were conducted using an industrial robot equipped with a pneumatic grinder and a force sensor. The models relating the commanded depth of cut to the corresponding normal grinding force was developed based on nine sets of experimental data. These data were collected at three different grinding paths with three different pseudo-random binary sequence amplitudes in order to fully investigate the dynamics of the process. Using the extended-least squares method for recursive parameter estimation, a linear fourth-order autoregressive moving average with exogenous variable model was successfully developed for the robotic disc-grinding process.