Towards a Standards-Based Methodology for Extending Manufacturing Process Models for Sustainability Assessment

Industrial use of natural resources are increasing at an alarming rate. Engineering and decision support tools are needed for analyzing and curbing industrial consumption of resources. Further, assessment methods to measure and indicate continuous improvement are also needed. Modeling individual manufacturing processes facilitates the generation of quantifiable evidence that improvements are being made. Such a modeling approach is developed and demonstrated in this paper to characterize sustainability performance of two metals inspection processes: magnetic particle inspection for ferromagnetic alloys and penetrant inspection for non-ferromagnetic alloys. Individual unit manufacturing process (UMP) models were developed by observing the inspection practices at an aircraft component manufacturer, and a mathematical basis for comparison with other inspection processes was identified. The paper further demonstrates the aggregation of performance metrics from all UMPs across a manufacturing process flow thus providing a basis for generating detailed sustainability performance assessments of manufactured products. By developing and documenting a comprehensive set of UMP models, more complete knowledge of manufacturing processes can be gained by industry practitioners, leading to continuous improvement of sustainability performance.

: Manufacturing process models are developed for a data-driven, reconfiguarable manufacturing facility. A human assembly process model is developed to characterize assembly processes. Machine and tool resources are characterized for manufacturing processes to provide a broad characterization of manufacturing process coverage in a hypothetical final assembly factory for infantry fighting vehicles. A sample population of manufacturing process parameters and resources is collected and incorporated into a manufacturing capability and process model library. Requirements for a virtual manufacturing environment are developed, and an initial, prototype version is implemented to facilitate process model validation, configuration, and experimentation.

This paper describes an on going research by the authors who are seeking methods to integrate an earlier developed manufacturing process modeling method in the virtual manufacturing (VM) technology. The primary aim is to attach special manufacturability evaluation to modeling of mechanical products. A more or less generic process model is generated for a cluster of manufacturing tasks then manufacturing process entities are created by evaluation of the manufacturing process model using part, form feature, production environment, production planning and cost information. Petri net manufacturing process model representation is used. It has proved to be a versatile virtual manufacturing tool in manufacturing process planning, evaluation of manufacturability and in manufacturing environment design. A process model evaluation procedure is emphasized in the course of which feasibility and resource requirements of the manufacturing are revealed. Firstly, techniques of evaluation of manufacturability, the applied approach to integration of shape and manufacturing process modeling and the related virtual manufacturing concepts are introduced. The process model is outlined and earlier relevant publications of the authors are cited. Then problems and solutions at evaluating of process model in close connection with design and environmental changes are detailed. Following this, integration of the method into a comprehensive virtual manufacturing concept is explained. Finally, main issues and future research plans are concluded.

Increasing complexity, and requirements for the precise creation of parts, necessitate the use of computer numerical control (CNC) manufacturing. This process involves programmed instructions to remove material from a workpiece through operations such as milling, turning, and drilling. This manufacturing technique incorporates various process parameters (e.g., tools, spindle speed, feed rate, cut depth), leading to a highly complex operation. Additionally, interacting phenomena between the workpiece, tools, and environmental conditions further add to complexity which can lead to defects and poor product quality. Two main areas are of focus for an efficient automated system: monitoring and swift quality assessment. Within these areas, the critical aspects ascertaining the quality of a CNC manufacturing operation are: 1) Tool wear: the inherent deterioration of machine components caused by prolonged utilization, 2) Chatter: vibration that occurs during the machining process, and 3) Surface finish: the final product’s surface roughness. Many research domains tend to focus on just one of these areas while neglecting the interconnected influences of all three. Therefore, to capture a more holistic and comprehensive assessment of a manufacturing process, the overall product quality should be considered, as that’s what ultimately counts. The integration of CNC systems with in-situ monitoring devices such as acoustic sensors, high-speed cameras, and thermal cameras is aimed at understanding the underlying physical aspects of the CNC machining process, including tool wear, chatter, and surface roughness. The incorporation of these monitoring devices has allowed the use of artificial intelligence and machine learning (ML) in smart CNC systems with hopes of increasing productivity, minimizing downtime, and ensuring product quality. By capturing the underlying phenomena that occur during the manufacturing process, users hope to understand the interlinking dynamics for zero-defect automated manufacturing. However, even though the use of ML methods has yielded noteworthy results in analyzing in-situ process data for CNC manufacturing, the black-box nature of these models and their tendency to focus predominantly on single-task objectives rather than multi-task scenarios pose challenges. In real-world part creation and manufacturing scenarios, there is often a need to address multiple interconnected tasks simultaneously which demands models that can multitask effectively. Yet, many ML models designed and trained for singular objectives are limited in their applicability and efficiency in more complex multi-faceted environments. Addressing these challenges, we introduce MTaskHD, a novel multi-task framework, that leverages hyperdimensional computing (HDC) to effortlessly fuse data from various channels and process signals while characterizing quality within a multi-task manufacturing operation. Moreover, it yields interpretable outcomes, allowing users to understand the process behind predictions. In a real-world experiment conducted on a hybrid 5-axis CNC Deckel-Maho-Gildemeister, MTaskHD was implemented to forecast the quality of three distinct features: left 25.4 mm counterbore diameter, right 25.4 mm counterbore diameter, and 2.54 mm milled radius. Demonstrating remarkable performance, the model excelled in predicting the quality levels of all three features in its multi-task configuration with an F1-Score of 95.3%, outperforming alternative machine learning approaches, including support vector machines, Naïve Bayes, multi-layer perceptron, convolutional neural network, and time-LeNet. The inherent multi-task capability, robustness, and interpretability of HDC collectively offer a solution for comprehending intricate manufacturing dynamics and operations.

In this paper, the growing significance of data analysis in manufacturing environments is exemplified through a review of relevant literature and a generic framework to aid the ease of adoption of regression-based supervised learning in manufacturing environments. To validate the practicality of the framework, several regression learning techniques are applied to an open-source multi-stage continuous-flow manufacturing process data set to typify inference-driven decision-making that informs the selection of regression learning methods for adoption in real-world manufacturing environments. The investigated regression learning techniques are evaluated in terms of their training time, prediction speed, predictive accuracy (R-squared value), and mean squared error. In terms of training time (TT), k-NN20 (k-Nearest Neighbour with 20 neighbors) ranks first with average and median values of 4.8 ms and 4.9 ms, and 4.2 ms and 4.3 ms, respectively, for the first stage and second stage of the predictive modeling of the multi-stage continuous-flow manufacturing process, respectively, over 50 independent runs. In terms of prediction speed (PS), DTR (decision tree regressor) ranks first with average and median values of 5.6784×106 observations per second (ob/s) and 4.8691×106 observations per second (ob/s), and 4.9929×106 observations per second (ob/s) and 5.8806×106 observations per second (ob/s), respectively, for the first stage and second stage of the predictive modeling of the multi-stage continuous-flow manufacturing process, respectively, over 50 independent runs. In terms of R-squared value (R2), BR (bagging regressor) ranks first with average and median values of 0.728 and 0.728, respectively, over 50 independent runs, for the first stage of the predictive modeling of the multi-stage continuous-flow manufacturing process, and RFR (random forest regressor) ranks first with average and median values of 0.746 and 0.746, respectively, over 50 independent runs, for the second stage of the predictive modeling of the multi-stage continuous-flow manufacturing process. In terms of mean squared error (MSE), BR (bagging regressor) ranks first with average and median values of 2.7 and 2.7, respectively, over 50 independent runs, for the first stage of the predictive modeling of the multi-stage continuous-flow manufacturing process, and RFR (random forest regressor) ranks first with average and median values of 3.5 and 3.5, respectively, over 50 independent runs, for the second stage of the predictive modeling of the multi-stage continuous-flow manufacturing process. All methods are further ranked inferentially using the statistics of their performance metrics to identify the best method(s) for the first and second stages of the predictive modeling of the multi-stage continuous-flow manufacturing process. A Wilcoxon rank sum test is then used to statistically verify the inference-based rankings. DTR and k-NN20 have been identified as the most suitable regression learning techniques given the multi-stage continuous-flow manufacturing process data used for experimentation.

Part machining is a discrete manufacturing process. In order to evaluate the manufacturing process, an intelligent modeling method based on the first-order predicate logic is proposed. First, the basic predicate formula is defined according to the machining method, and the predicate and variables are illustrated in detail. Thus, the process representation is completed. Second, to construct the process model, the modeling element is put forward, which includes three nodes. Components of modeling element are, respectively, discussed, as well as the mapping relationship between modeling element and predicate. After the definition of modeling predicate formula, five basic inference rules are established. Consequently, the manufacturing process model is constructed. Third, on the basis of the process model, the process simulation is carried out to evaluate the manufacturing performances, such as the production efficiency, the utilization rate of machining equipment, the production bottleneck, etc. Finally, a case study is conducted to explain this modeling method.

Digital twins for electro-physical, chemical, and photonic processes

Fuzzy regression is a commonly used approach for modeling manufacturing processes in which the availability of experimental data is limited. Fuzzy regression can address fuzzy nature of experimental data in which fuzziness is not avoidable while carrying experiments. However, fuzzy regression can only address linearity in manufacturing process systems, but nonlinearity, which is unavoidable in the process, cannot be addressed. In this paper, an evolutionary fuzzy regression which integrates the mechanism of a fuzzy regression and genetic programming is proposed to generate manufacturing process models. It intends to overcome the deficiency of the fuzzy regression, which cannot address nonlinearities in manufacturing processes. The evolutionary fuzzy regression uses genetic programming to generate the structural form of the manufacturing process model based on tree representation which can address both linearity and nonlinearities in manufacturing processes. Then it uses a fuzzy regression to determine outliers in experimental data sets. By using experimental data excluding the outliers, the fuzzy regression can determine fuzzy coefficients which indicate the contribution and fuzziness of each term in the structural form of the manufacturing process model. To evaluate the effectiveness of the evolutionary fuzzy regression, a case study regarding modeling of epoxy dispensing process is carried out.

As the need for quickly exploring a textile manufacturing process is increasingly costly along with the complexity in the process. The development of manufacturing process modeling has attracted growing attention from the textile industry. More and more researchers shift their attention from classic methods to the intelligent techniques for process modeling as the traditional ones can hardly depict the intricate relationships of numerous process factors and performances. In this study, the literature investigating the process modeling of textile manufacturing is systematically reviewed. The structure of this paper is in line with the procedure of textile processes from yarn to fabrics, and then to garments. The analysis and discussion of the previous studies are conducted on different applications in different processes. The factors and performance properties considered in process modeling are collected in comparison. In terms of inputs’ relative importance, feature selection, modeling techniques, data distribution, and performance estimations, the considerations of the previous studies are analyzed and summarized. It is also concluded the limitations, challenges, and future perspectives in this issue on the basis of the summaries of more than 130 related articles from the point of views of textile engineering and artificial intelligence.

The positioning accuracy of computer numerical control (CNC) machine tools is mainly limited by the manufacturing accuracy of their linear and circular motion axes and by the long-term dimensional stability of their structures. Maximizing this accuracy can prove to be a particularly challenging task, especially for large-sized systems. In fact, heat-induced deformations, long-period deformation of foundations and the manufacturing process itself, these all cause time-dependent structural deformations of the machine body, which are difficult to model and to predict. The usual approach is a model-based prediction of structural deformations, which is followed by a compensation of positioning errors at CNC level. This approach is often limited by the complexity of the problem from both geometrical (system geometry can be very complex and it can vary in time) and physical (it is difficult to model and consider any possible load type and loading condition) point of view. As a consequence, only limited success has been achieved in active error compensation based on the modelling of the relationship between the generalized dynamic loads and the structural deformation field. This paper illustrates a different approach in active error compensation, which exploits a new measurement system able to provide real-time measurement of the displacement field of a given structural component, without any model about its dynamic/thermal structural behavior.

It has been shown that a manufacturing process can be modeled (learned) using Multi-Layer Perceptron (MLP) neural network and then optimized directly using the learned network. This paper extends the previous work by examining several different MLP training algorithms for manufacturing process modeling and three methods for process optimization. The transformation method is used to convert a constrained objective function into an unconstrained one, which is then used as the error function in the process optimization stage. The simulation results indicate that: (i) the conjugate gradient algorithms with backtracking line search outperform the standard BP algorithm in convergence speed; (ii) the neural network approaches could yield more accurate process models than the regression method; (iii) the BP with simulated annealing method is the most reliable optimization method to generate the best optimal solution, and (iv) process optimization directly performed on the neural network is possible but cannot be especially automated totally, especially when the process concerned is a mixed integer problem.

This paper proposes fuzzy Colored Petri Nets with stochastic time delay (FCPN-std) as one of methodologies of incorporating qualitative uncertainties and stochastic information in a manufacturing system. As a combined model with fuzzy colored Petri Nets (FCPNs) and stochastic Petri Nets (SPNs), it captures and models statistical and qualitative uncertainties in manufacturing processes. A parametric modular design methodology is proposed for effective design of complex processes. A mold producing process in digital camera making industry is modeled as an example application. The model incorporating uncertain and stochastic information is represented using extended Petri Net Markup Language (xPNML) architecture. xPNML formatted fuzzy colored Petri Net with stochastic time delay is shown as an effective design methodology and representation tool for describing manufacturing processes and capturing uncertainties.

Digital manufacturing techniques, such as computer numerically controlled (CNC) machining and additive manufacturing (AM), have greatly increased the design space available to engineers with respect to both geometry and materials. Obviously, certain geometries cannot be realized by anything other than a layer-by-layer AM process, and certain materials can only be formed using traditional subtractive approaches, leaving designers with little choice. However, for the wide array of structures that could be realized in a variety of different ways, the choice of manufacturing process is just another faction in the overall design process. Planar lattice structures have complex repeating geometric features which are best designed using computer aided design (CAD) tools, and these structures can easily be manufactured using a variety of different digital manufacturing processes. For this study a set of planar lattice structures were designed and fabricated from acrylonitrile butadiene styrene (ABS) plastic using both a CNC milling machine and a material extrusion AM machine. All test lattices were then loaded in tension to failure to determine their mechanical performance. The determined properties were then compared to analytical predictions from the literature, and an analysis of the impact of a chosen manufacturing process was performed. The results obtained show good agreement with publish models for lattice performance, and additionally suggest that the choice of manufacturing process has a measurable effect on the mechanical performance of the resulting structures for this specific material and geometry combination, but it is important to note that this study is just an example of how to mitigate the uncertainties associated with the freedoms offered by advanced manufacturing processes, and more work in needed in this important area of study.

Traditional Computer Numerical Control (CNC) machines use ISO6983 (G/M code) for part programming. G/M code has a number of drawbacks and one of them is lack of interoperability. The Standard for the Exchange of Product for NC (STEP-NC) as a potential replacement for G/M code aims to provide a unified and interoperable data model for CNC. In a modern CNC machine tool, more and more motors, actuators and sensors are implemented and connected to the NC system, which leads to large quantity of data being transmitted. The real-time Ethernet field-bus is faster and more deterministic and can fulfill the requirement of data transmission in the high-speed and high-precision machining scenarios. It can provide more determinism on communication, openness, interoperability and reliability than a traditional field-bus. With a traditional CNC system using G/M code, when the machining is interrupted by incidents, restarting the machining process is time-consuming and highly experience-dependent. The proposed CNC controller can generate just-in-time tool paths for feature-based machining from a STEP-NC file. When machining stoppage occurs, the system can recover from stoppage incidents with minimum human intervention. This is done by generating new tool paths for the remaining machining process with or without the availability of the original cutting tool. The system uses a real-time Ethernet field-bus as the connection between the controller and the motors.

This paper presents the most important aspects of semiconductor processes modeling and a process/device optimization technique based on neural network models. Efficiency of adaptive neural networks in complex process modeling and inverse modeling is demonstrated through modeling of ion implantation process. Also, efficiency of Technology Computer-Aided Design (TCAD) system containing direct and inverse neural network models is demonstrated through optimization of complex manufacturing process flow of low voltage power VDMOSFET. Results of optimization obtained by using direct and inverse neural network models are compared with those obtained by using world-known process simulator MUSIC 2 and modified version of device simulator MINIMOS 6, and an excellent agreement is achieved. It is shown that incorporation of neural network models in TCAD systems significantly improves optimization technique, thus leading to significant reduction of computational time.