Desai, Kaushal A

In-process dimension measurement is critical to achieving higher productivity and realizing smart manufacturing goals during machining operations. Vision-based systems have significant potential to serve for in-process dimensions measurements, reduce human interventions, and achieve manufacturing-inspection integration. This paper presents early research on developing a vision-based system for in-process dimension measurement of machined cylindrical components utilizing image-processing techniques. The challenges with in-process dimension measurement are addressed by combining a deep learning-based object detection model, You Only Look Once version 2 (YOLOv2), and image processing algorithms for object localization, segmentation, and spatial pixel estimation. An automated image pixel calibration approach is incorporated to improve algorithm robustness. The image acquisition hardware and the real-time image processing framework are integrated to demonstrate the working of the proposed system by considering a case study of in-process stepped shaft diameter measurement. The system implementation on a manual lathe demonstrated robust utilities, eliminating the need for manual intermittent measurements, digitized in-process component dimensions, and improved machining productivity.

Manual assessment of laboratory exercises and assignments by human instructors is subjective and time-consuming, introducing errors, inconsistencies, and biases. A preliminary study developing an automated evaluation system utilizing computer vision has been proposed in this article to address these challenges. It utilizes computer vision to assess the accuracy and quality of components machined by students enrolled during the manufacturing laboratory course. The system includes image acquisition hardware, algorithms for objective decisions, and an interface for evaluating student performance. The system is implemented during one of the laboratory classes, and a comparative assessment is carried out with manual evaluations followed by student feedback. It has been shown that the system can address subjectivity concerns, reducing the workload on course instructors and teaching assistants. The research also broadens the utility of computer vision in manufacturing education and creates interest among enrolled students to appreciate the role of newer technologies in a core discipline.

Machine vision systems commonly utilize Convolutional Neural Networks (CNNs) for in-line surface defect detection of manufactured components. The prediction abilities of vision-based inspection systems deteriorate with time as the defect detection model trained on fixed image datasets fails to accommodate deviations. This paper proposes a human-guided progressive learning approach that systematically imparts learning of new features to the CNN-powered vision-based defect detection system. The approach augments the surface defect detection model with human intelligence, using an intuitive user interface to address model drift. The human expert monitors the trained model performance under specific conditions leading to the change of characteristics during implementation, identifies misclassifications, and initiates re-training. The algorithm accumulates misclassified data till a pre-defined threshold level is reached or a human expert terminates inspection. The misclassified results merge with the original datasets for progressive re-training using a strategy similar to the base model development. The present work utilizes pre-trained CNN Efficientnet-b0 to develop the surface defect detection model for tapered roller inspection through transfer learning. It is concluded that the progressive re-training improves defect detection performance and reduces misclassifications. The Matthews Correlation Coefficient (MCC) score, derived from the confusion matrix, showed improvement from 0.6 to 0.82 after four iterations. A cross-model benchmarking study is also performed to show the versatility of the proposed approach. The present work demonstrated that the human-guided progressive learning approach can provide adaptability to vision-based surface defect detection utilizing deep learning algorithms and enhance system performance during real-time implementation.

Identifying tool wear state is essential for machine operators as it assists in informed decisions for timely tool replacement and subsequent machining operations. As each wear state corresponds to a unique mitigation strategy, timely identification is vital while implementing solutions to minimize tool wear. The paper presents a novel Human Guided-eXplainable Artificial Intelligence (HG-XAI) approach for identifying the tool wear state by integrating human intelligence and eXplainable AI with a pre-trained Convolutional Neural Network (CNN), Efficient-Net-b0 model. The tool wear states were identified based on different wear mechanisms during the machining of IN718. The study considers four distinct tool wear states, i.e., Flank, Flank+BUE, Flank+Face, and Chipping, representing abrasion, adhesion, diffusion, and fracture wear mechanisms. The image-based datasets were created to depict various tool wear states by machining IN718 at varying surface speeds. The effectiveness of the proposed HG-XAI approach was evaluated by comparing its prediction accuracy with a standalone Efficient-Net-b0 model lacking human intelligence and XAI. Further, the scalability of the HG-XAI approach was examined by predicting wear states from images acquired at different cutting parameters. The results from the present study showed that the HG-XAI approach can predict the tool wear state with an accuracy of 93.08% and is scalable to variations in cutting conditions. Also, the proposed approach can be extended while developing vision-based on-machine tool wear monitoring systems.

The existing 3-D printing techniques have several disadvantages such as aliasing and difficulty in building around inserts due to limited motions associated with the equipment. The limitation of build direction results in poor surface finish due to aliasing (or layer stair-stepping) and adverse material properties in certain directions which limits use of 3-D printing for many industrial applications. The present study investigates the application of Parallel Kinematic Machines (PKMs) in achieving multidirectional 3-D printing. The proposed architecture addresses some of the limitations of existing Fused Deposition Modelling (FDM)-based 3-D printer by allowing six-axis motions between extruder and platform while building the component. The study explores the application of Stewart–Gough Platform (SGP) further for 3-D printing and illustrates its capability as a viable solution for multi-axis FDM. The design of SGP for multidirectional FDM is realized for optimal dexterity using bulk dexterity index. The study discusses details of the optimization formulation and consequent results associated with the same. A conceptual design of the SGP is subsequently proposed based on the results of the optimization. The proposed SGP-based machine architecture is expected to offer advantages such as improved surface finish and control of directional properties, which signifies push towards freeform fabrication using multidirectional 3-D printing.

Micro-end milling is used for manufacturing of complex miniaturized components precisely in wide range of materials. It is important to predict cutting forces accurately as it plays vital role in controlling tool and workpiece deflections as well as tool wear and breakage. The present study attempts to incorporate process characteristics such as edge radius of cutting tool, effective rake and clearance angles, minimum chip thickness, and elastic recovery of work material collectively while predicting cutting forces using mechanistic model. To incorporate these process characteristics effectively, it is proposed to divide cutting zone into two regions: shearing- and ploughing-dominant regions. The methodology estimates cutting forces in each partitioned zone separately and then combines the same to obtain total cutting force at a given cutter rotation angle. The results of proposed model are validated by performing machining experiments over a wide range of cutting conditions. The paper also highlights the importance of incorporating elastic recovery of work material and effective rake and clearance angle while predicting cutting forces. It has been observed that the proposed methodology predicts the magnitude and profile of cutting forces accurately for micro-end milling operation.

Micro-milling has shown great potential in producing complex miniaturized components over wide range of materials. It can also fabricate micro-products in small batches efficiently and economically. In spite of these advantages, several challenges hinder its ability to produce components with better dimensional accuracy. Among several factors, tool deflection is one of the major sources of surface error on machined parts and features. Therefore, it is necessary to develop accurate and reliable process models to analyze and improve performance of the process. This study presents a methodology to determine cutting forces and surface error in the presence of tool deflections for micro-milling operation. Tool deflections have considerable influence on instantaneous uncut chip thickness. As tool deflection alters tooth trajectories and instantaneous uncut chip thickness, the rigid cutting force model needs to be modified suitably to consider the effect of deflections. This aspect has been incorporated in the model by modifying tool center location and tooth trajectories iteratively. The convergence of an iterative algorithm determining stable chip thickness is obtained by comparing RMS deviation of average chip thickness between two successive tooth passes. The axial variation of surface error due to tool deflections is estimated using surface generation mechanism. The proposed model is implemented in the form of a computational program to predict cutting force and surface error. The results of computational model are substantiated further by conducting machining experiments. It is shown that the proposed model predicts cutting forces fairly well in the presence of tool deflections. A comparison between predicted variation of surface error and 3D images of machined surface captured using optical microscope showed good qualitative agreement in the error profiles.

End milling of the curved surfaces is characterized by significant amount of engagement variation along the tool path which results in deflection-induced surface error on the machined components. Feed rate regulation cannot be used in this case to minimize the surface errors as it necessitates continuous change in feed rate along the tool path which deteriorates surface finish and lowers productivity. Another approach in the form of tool path modification scheme was also proposed in the literature to minimize engagement variation at corners of a pocket. This approach has been effective in reducing fluctuations of engagement at corners and cannot be applied directly in machining of curved geometries where engagement variation is continuous along the entire tool path. Keeping these limitations in focus, a new methodology is proposed in this article to minimize surface error variation that modifies machining strategy for the curved components. The strategy proposes machining of the curved components into two different stages: first, generating the modified semi-finishing geometry using algorithm proposed in this article and then producing the desired geometry using conventional contour parallel tool path. The methodology generates the modified semi-finishing geometry such that the engagement offered is constant when a finishing pass is made. The effectiveness of the proposed methodology has been verified by performing computational studies and machining experiments on typical curved geometries. It is observed that the proposed methodology is quite effective in minimizing the variation in the surface error on the machined components.

The recent industrial transformations aim to augment Information and Communication Technologies (ICT) with manufacturing equipment for prognosis, predictive maintenance, and enhanced human–machine interactions. Integrating ICT with legacy manual machines is challenging as human operators are solely responsible for operation monitoring and control. This paper presents a smart system that complements manual drilling machine operators in identifying tool wear states and chatter onset. The system also assists in twist drill replacement or regrinding decisions and achieving stable operating conditions. The in-process twist drill wear state and chatter onset are identified by integrating accelerometer and acoustic emission sensors with decision-making algorithms and presenting the status using Human Machine Interface (HMI) devices. The Root Mean Square (RMS) and Support Vector Machine (SVM) algorithms extract features and in-process drilling operation status from the sensor data. The SVM is trained by performing a set of drilling experiments and utilizing the expertise of skilled operators. The decision-making model and sensors are integrated with a legacy manual drilling machine using an HMI device to display tool wear state and chatter onset, thereby improving operator perceptions while performing operations. The system performance is corroborated by conducting experiments for various twist drill-work material combinations. It is concluded that the developed system can effectively capture in-process tool wear and stability states. The proposed system can be implemented as a potential solution for the guided operation and monitoring of legacy manual drilling machines. © The Author(s), under exclusive licence to Springer-Verlag London Ltd., part of Springer Nature 2024.