Mouhab Meshreki

Dynamics Modeling and Analysis of Thin-Walled Aerospace Structures for Fixture Design in Multiaxis Milling

Milling of thin-walled aerospace structures is a critical process due to the high flexibility of the workpiece. Current practices in the fixture design and the choice of cutting parameters rely solely on conservative guidelines and the designers experience. This is a result of the lack of computationally efficient dynamic models to represent the dynamic response of the workpiece during machining, and the interaction between the workpiece, fixture and the cutting forces. This paper presents a novel dynamic formulation of typical thin-walled pockets encountered in aerospace structures. It is based on an analytical description of a five-sided pocket using a plate model. An off-line calibration of the model parameters, using global and local optimization, is performed in order to match the dynamic response of the pocket structure. The developed simplified model is based on Rayleighs energy method. Various pocket shapes are examined under different loading conditions and compared to finite element (FE) predictions and experimental results. In both cases, the results obtained by the developed model are in excellent agreement. This proposed approach resulted in one to two orders of magnitude reduction in computational time when compared to FE models, with a prediction error less than 10%.

Development of a New Model for the Varying Dynamics of Flexible Pocket-Structures During Machining

Many of the aerospace components are characterized by having pocket-shaped thin-walled structures. During milling, the varying dynamics of the workpiece due to the change of thickness affects the final part quality. Available dynamic models rely on computationally prohibitive techniques that limit their use in the aerospace industry. In this paper, a new dynamic model was developed to predict the vibrations of thin-walled pocket structures during milling while taking into account the continuous change of thickness. The model is based on representing the change of thickness of a pocket-structure with a two-directional multispan plate. For the model formulation, the Rayleigh–Ritz method is used together with multispan beam models for the trial functions in both the x- and y-directions. An extensive finite element (FE) validation of the developed model was performed for different aspect ratios of rectangular and nonrectangular pockets and various change of thickness schemes. It was shown that the proposed model can accurately capture the dynamic effect of the change of thickness with prediction errors of less than 5% and at least 20 times reduction in the computation time. Experimental validation of the models was performed through the machining of thin-walled components. The predictions of the developed models were found to be in excellent agreement with the measured dynamic responses.

A New Analytical Formulation for the Dynamics of Multipocket Thin-Walled Structures Considering the Fixture Constraints

Milling of thin-walled aerospace structures is a critical and challenging process. Available models for the prediction of the effect of the fixture on the dynamic response of flexible workpieces are computationally demanding and fail to represent practical cases for milling of thin-walled structures. Based on the analysis of typical structural components encountered in the aerospace industry, a generalized unit-element, with the shape of an asymmetric pocket, was identified to represent the dynamic response of these components. Accordingly, a computationally efficient dynamic model was developed to predict the dynamic response of typical thin-walled aerospace structures using the Rayleigh-Ritz method. In the formulation of this model, the dynamics of a 3D pocket is represented by an equivalent 2D multispan plate taking into account the effect of deformable fixture supports. The developed model was validated numerically and experimentally for different workpiece geometries and various types of loading. This model resulted in one to two orders of magnitude reduction in computation time when compared with the finite element models, with prediction errors less than 10%. The developed model meets the conflicting requirements of prediction accuracy and computational efficiency needed for interactive fixture design.

Monitoring and Control of Machining Process by Data Mining and Pattern Recognition

In this paper we present a novel approach to the problem of understanding, monitoring, and controlling the machining process of composites materials. The approach is called Logical Analysis of Data (LAD). It is based on data mining and pattern recognition, and uses a machine learning artificial intelligence technique. This novel approach is used for the first time in order to define machining conditions that lead to conforming products, and also conditions which will lead to nonconforming products. In this paper, we introduce the LAD technique, we apply it to the machining of composites, and we report on the results based on data obtained experimentally. We conclude with a discussion of the potential use of LAD in manufacturing.

Process control based on pattern recognition for routing carbon fiber reinforced polymer

Carbon fiber reinforced polymer (CFRP) is an important composite material. It has many applications in aerospace and automotive fields. The little information available about the machining process of this material, specifically when routing process is considered, makes the process control quite difficult. In this paper, we propose a new process control technique and we apply it to the routing process for that important material. The measured machining conditions are used to evaluate the quality and the geometric profile of the machined part. The machining conditions, whether controllable or uncontrollable are used to control part accuracy and its quality. We present a pattern-based machine learning approach in order to detect the characteristic patterns, and use them to control the quality of a machined part at specific range. The approach is called logical analysis of data (LAD). LAD finds the characteristic patterns which lead to conforming products and those that lead to nonconforming products. As an example, LAD is used for online control of a simulated routing process of CFRP. We introduce the LAD technique, we apply it to the high speed routing of woven carbon fiber reinforced epoxy, and we compare the accuracy of LAD to that of an artificial neural network, since the latter is the most known machine learning technique. By using experimental results, we show how LAD is used to control the routing process by tuning autonomously the routing conditions. We conclude with a discussion of the potential use of LAD in manufacturing.

Effect of Tool Kinematics on the Drilling Forces and Temperature in Low Frequency High Amplitude Vibration Assisted Drilling

Defects associated with drilling of fiber reinforced polymers (FRPs) are of major economic and safety concerns for aerospace manufacturers. Delamination of layers and thermal damage of the matrix are the most critical defects associated with drilling of FRP laminates, which can be avoided by keeping the drilling forces and temperatures below some threshold levels. Vibration-assisted drilling (VAD) is an emerging drilling process that uses intermittent cutting to reduce the drilling forces and temperatures, and achieve easier chip removal compared to conventional drilling. In this paper an extensive experimental study has been conducted to provide insight into the effect of the tool kinematics corresponding to the VAD parameters (speed, feed, frequency and amplitude) on the geometry of the formed chip determined by the intersection of the trajectories of the cutting edges as well as on the drilling forces and temperature. The combinations of the VAD parameters used in this study were selected from ranges of speeds 6,000 rpm to 12,000 rpm, feeds 0.05 mm/rev to 0.15 mm/rev, frequencies 30 Hz and 60 Hz, and amplitudes 40 μm to 400 μm. The Amplitude and feed were found to have the most dominant effect on the VAD forces, while the feed and speed had the dominant effect on the VAD temperatures. The thermal performance of the VAD process was found to be enhanced by the formation of vortices in the air gap created by the separation between the tool and the machined surface, which is mainly controlled by the feed and the rotational speed of the tool.Copyright

Diagnosis of machining outcomes based on machine learning with Logical Analysis of Data

Force is considered to be one of the indicators that best describe the machining process. Measured force can be used to evaluate the quality and geometric profile of the machined part. In this paper, a combinatorial optimization approach is used to characterize the effect of force on the quality of a machined part made of Carbon Fiber Reinforced Polymers (CFRP) material. The approach is called Logical Analysis of Data (LAD) and is based on machine learning and pattern recognition. LAD is used in order to map the machining conditions, in terms of force and torque that lead to conforming products and those which lead to nonconforming products. In this paper, the LAD technique is applied to the drilling of CFRP plates, and the results, based on data obtained experimentally, are reported. A discussion of the potential use of LAD in manufacturing concludes the paper.

Optimization of Cutting Conditions in Vibration Assisted Drilling of Composites via a Multi-Objective EGO Implementation

A recent and promising technique to overcome the challenges of conventional drilling is vibration-assisted drilling (VAD) whereby a controlled harmonic motion is superimposed over the principal drilling feed motion in order to create an intermittent cutting state. Two additional variables other than the feed and the speed are introduced, namely the frequency and the amplitude of the imposed vibrations. Optimum selection of cutting conditions in VAD operations of composite materials is a challenging task due to several reasons; such as the increase in the number of controllable variables, the need for costly experimentation, and the limitation on the number of experiments that can be performed before tool degradation becomes an issue in the reliability of measurements. Additionally, there are often several objectives to consider, some of which may be conflicting, while others may be somewhat correlated. Pareto-optimality analysis is needed for conflicting objectives; however the existence of several objectives (high-dimension Pareto space) makes the generation and interpretation of Pareto solutions difficult. An attractive approach to the optimization task is thus to employ Kriging meta-models in a multi-objective efficient global optimization (m-EGO) framework for incremental experimentation of optimal setting of the cutting parameters. Additional challenge posed by constraints on machine capabilities is accounted for through domain transformation of the design variables prior to the construction of the Kriging models. Study results using a baseline exhaustive experimental data shows opportunity for employing m-EGO for the generation of well distributed Pareto-frontiers with fewer experiments.Copyright

Multi-Objective Selection of Cutting Conditions in Advanced Machining Processes via an Efficient Global Optimization Approach

Optimum selection of cutting conditions in high-speed and ultra-precision machining processes often poses a challenging task due to several reasons; such as the need for costly experimental setup and the limitation on the number of experiments that can be performed before tool degradation starts becoming a source of noise in the readings. Moreover, oftentimes there are several objectives to consider, some of which may be conflicting, while others may be somewhat correlated. Pareto-optimality analysis is needed for conflicting objectives; however the existence of several objectives (high-dimension Pareto space) makes the generation and interpretation of Pareto solutions difficult. The approach adopted in this paper is a modified multi-objective efficient global optimization (m-EGO). In m-EGO, sample data points from experiments are used to construct Kriging meta-models, which act as predictors for the performance objectives. Evolutionary multi-objective optimization is then conducted to spread a population of new candidate experiments towards the zones of search space that are predicted by the Kriging models to have favorable performance, as well as zones that are under-explored. New experiments are then used to update the Kriging models, and the process is repeated until termination criteria are met. Handling a large number of objectives is improved via a special selection operator based on principle component analysis (PCA) within the evolutionary optimization. PCA is used to automatically detect correlations among objectives and perform the selection within a reduced space in order to achieve a better distribution of experimental sample points on the Pareto frontier. Case studies show favorable results in ultra-precision diamond turning of Aluminum alloy as well as high-speed drilling of woven composites.Copyright

Experimental Characterization and Multi-Objective Optimization of the Orbital Drilling Process of CFRP

Defects associated with drilling of Carbon Fiber-Reinforced Polymers (CFRPs) are of major economic and safety concerns for aerospace manufacturers. One of the most critical defects associated with drilling of CFRP laminates is delamination of layers which can be avoided by keeping the drilling forces below some threshold levels. Orbital Drilling (OD) is an emerging drilling process that exhibits lower cutting forces and temperatures, easier chip removal, higher produced surface quality, longer tool life, and a high possibility for dry machining. The OD process is featured by cyclic engagement and disengagement between the tool and the workpiece whereby a considerable part of the work done by the tool is directed towards the tangential direction while the work done in the axial direction is reduced. This reduces the risk of delamination at the exit. The objective of this research work is to investigate the effect of the OD process key parameters with respect to the produced hole attributes (surface roughness, delamination, and hole accuracy), as well as the cutting forces and temperatures. All the OD tests were performed under dry conditions using a four-flute 6.35 mm end-mill. The cutting forces were recorded using a 3-component dynamometer Kistler 9255B and cutting temperatures were measured using a FLIR ThermoVision A20M Infrared camera at the holes exit. A full factorial design of the experiment was used whereby the feeds varied from 60 to 360 mm/min and the speeds from 6,000 to 16,000 rpm. The test material used was a quasi-isotropic laminate comprising woven graphite epoxy prepreg. Analysis of the results showed 45% reduction in the axial force component in orbital drilling (OD), compared to conventional drilling. None of the holes produced by the entire set of experiments has experienced any entry or exit delamination. ANOVA was used to identify the significance of the controllable variables on the experimental outputs. To overcome the challenge of optimizing the competing parameters of the hole quality attributes while maximizing the productivity, an algorithm was applied by hybridizing Kriging as a meta-modeling technique with evolutionary multi-objective optimization to optimize the cutting parameters.Copyright