Quality Deviation Analysis of a Multistage Machining Process

This article aims to compare three different approaches already proposed by the authors regarding the three-dimensional error stack-up analysis in machining. Although the developed error stack-up analysis method comprises of different manufacturing processes, this article is limited to machining operations. This article is built on the previous research of the authors namely the model of manufactured part (MMP) which models the different geometrical deviation 3D impacts on the geometrical and dimensional quality of part produced (error stack-up) in a multi-stage (MS) machining process. In this study, the MMP will be used for developing numerical solution techniques for the aim of tolerance analysis in a MS machining process. The developed solution techniques are classified in two main categories: worst case analysis (WCA) (worst part produced) and stochastic analysis (STA). This article provides an overview of the MMP model and the different solution techniques. The different strategies to simulate the deviation parameters of the model are then discussed. For each of the three proposed solution techniques, its advantages and disadvantages are explored in detail. The performance of the solution techniques is then compared from different points of view (i.e. rapidity, convergence to the global minimum and analysed case … ) through a 3D example, and finally some perspectives are presented.

With the increasing scale and complexity of system, it is very necessary to analyze the safety of complex system. Fault tree is an effective method to safety analysis. However, traditional fault tree relies on manual construction and analysis. When fault nodes and systems are complex, the efficiency and correctness of manual analysis can hardly be guaranteed. To the varied understanding of analysts, it is difficult to ensure the consistency of failure mode and system architecture due to the different understanding from safety analysts and system designers. The same node needs fault analysis again in different systems, which has poor reusability and low efficiency. AltaRica is a fault-oriented Safety Modeling Language. It takes the guard transformation system(GTS) as its core, describes nodes and faults with a style of reusable object-oriented language, and describes information of interaction between nodes and systems through interface connections between nodes and nested systems. Therefore, this paper proposed an automatic system modeling and fault analysis method and its detailed computer algorithm on single class node, multiple nodes and nodes with subsystems based on AltaRica. Finally we developed a software prototype and carry out the automatic modeling and fault analysis in a detailed example. The results showed that the proposed method, algorithm and software prototype can realize automatic graphical modeling of the system on AltaRica, the automatic fault analysis is correct and efficient, has reusability of modeling and fault analysis, and greatly improve the accuracy, objectivity and efficiency of fault modeling and analysis.

Vertically integrated cloud liquid water has been successfully analysed from both the Special Sensor Microwave Imager (SSM/I) and the Advanced Microwave Sounding Unit (AMSU). However microwave radiometers provide very little information on the vertical cloud liquid water profile. When considering the assimilation of microwave sounding channels, such as those on the AMSU or the Special Sensor Microwave Imager Sounder (SSMIS) the sensitivity of the measurements to cloud liquid water is dependent on the altitude of the cloud. Cloud well below the peak of the weighting function will tend to increase the measured brightness temperatures whereas cloud at or above the weighting function peak will tend to lower the measured brightness temperatures. Therefore if temperature and humidity sounding data are to be assimilated where cloud is present it will be necessary to analyse simultaneously the temperature, humidity and cloud profiles. One approach is to use a one dimensional variational analysis (1D-var) with a single moisture control variable, total water, where total water is the sum of water vapour, cloud liquid water, cloud ice water, rain and snow. For simplicity in this study rain and snow are neglected and ice cloud is assumed to be transparent. The analysis of cloud from the total water approach is compared with a variational analysis with separate cloud and water vapour control variables. The suitability of a total water control variable for variational analysis of cloudy AMSU, SSM/I or SSMIS data is then discussed.

In design of a multistage machining process, process tolerance allocation at each stage and design of process layouts, in particular, the fixture layouts, should be optimized considering dimensional tolerance stackup or variation propagation. When multiple error sources contribute to the tolerance stackup, the dimensionality of the design space could be large and design solution may not be unique. One strategy is to prioritize the allocation of tolerances to different error sources at each stage through proper selection of cost functions. Considering the fact that the cost function selection can be very subjective and the knowledge regarding cost structures is very limited, we propose a hierarchical process design method using error equivalence concept to aggregate multiple error sources together. The main idea is to allocate tolerances to the aggregated error sources at each stage and then further distribute the tolerance to individual error sources through cost analysis. The advantage is two-fold: (1) limiting the impact of cost function selection within individual stages to avoid overhaul of process design caused by subtle change in cost functions, and (2) enabling the optimization of fixture layouts to reduce the overall tolerance stackup due to multiple error sources. To reduce computational load in optimizing process layouts, a computer experiment model is developed to explore a large number of process design alternatives. The robustness of the optimal tolerance is evaluated through sensitivity analysis, which provides guidance for process design. We illustrate the error-equivalence-based process design method by a multistage machining process. The results have shown that the proposed method could significantly reduce the design space and increase the design efficiency.

Just-in-time defect prediction based on AST change embedding

Satisfying the quality requirement of products in multistage machining processes (MMPs) is significant and challenging nowadays. In spite of the success of the stream of variation (SoV) theory in variation propagation modelling for MMPs, the absence of elastic deformation variations could be an important factor that limits its application in variable stiffness structure (VSS) workpieces. To this end, a generic variation propagation framework in MMPs for VSS workpieces is established, incorporating the induction and propagation of elastic deformation variations. Region division strategy is adopted according to the characteristics of VSS workpieces, and the analytic solutions of elastic deformation in different regions are solved by contact theory and elastic theory. The effectiveness and accuracy of the proposed model are verified by a six-stage machining process on a four-cylinder engine block, and the proposed model is compared with the conventional SoV model resulting in a significant improvement on quality prediction.

In the traditional software defect prediction methodology, the historical record (dataset) of the same project is partitioned into training and testing data. In a practical situation where the project to be predicted is new, traditional software defect prediction cannot be employed. An alternative method is cross-project defect prediction, where the historical record of one project (source) is used to predict the defect status of another project (target). The cross-project defect prediction method solves the limitations of the historical records in the traditional software defect prediction method. However, the performance of cross-project defect prediction is relatively low because of the distribution differences between the source and target projects. Furthermore, the software defect dataset used for cross-project defect prediction is characterized by high-dimensional features, some of which are irrelevant and contribute to low performance. To resolve these two issues, this study proposes a transformation and feature selection approach to reduce the distribution difference and high-dimensional features in cross-project defect prediction. A comparative experiment was conducted on publicly available datasets from the AEEEM. Analysis of the results obtained shows that the proposed approach in conjugation with random forest as the classification model outperformed the other four state-of-the-art cross-project defect prediction methods based on the commonly used performance evaluation metric F1_score.

Recent studies of software defect prediction typically produce datasets, methods and frameworks which allow software engineers to focus on development activities in terms of defect-prone code, thereby improving software quality and making better use of resources. Many software defect prediction datasets, methods and frameworks are published disparate and complex, thus a comprehensive picture of the current state of defect prediction research that exists is missing. This literature review aims to identify and analyze the research trends, datasets, methods and frameworks used in software defect prediction research betweeen 2000 and 2013. Based on the defined inclusion and exclusion criteria, 71 software defect prediction studies published between January 2000 and December 2013 were remained and selected to be investigated further. This literature review has been undertaken as a systematic literature review. Systematic literature review is defined as a process of identifying, assessing, and interpreting all available research evidence with the purpose to provide answers for specific research questions. Analysis of the selected primary studies revealed that current software defect prediction research focuses on five topics and trends: estimation, association, classification, clustering and dataset analysis. The total distribution of defect prediction methods is as follows. 77.46% of the research studies are related to classification methods, 14.08% of the studies focused on estimation methods, and 1.41% of the studies concerned on clustering and association methods. In addition, 64.79% of the research studies used public datasets and 35.21% of the research studies used private datasets. Nineteen different methods have been applied to predict software defects. From the nineteen methods, seven most applied methods in software defect prediction are identified. Researchers proposed some techniques for improving the accuracy of machine learning classifier for software defect prediction by ensembling some machine learning methods, by using boosting algorithm, by adding feature selection and by using parameter optimization for some classifiers. The results of this research also identified three frameworks that are highly cited and therefore influential in the software defect prediction field. They are Menzies et al. Framework, Lessmann et al. Framework, and Song et al. Framework.

Fault prognosis technology, as the key to realize autonomy support and carry out condition-based maintenance (CBM), has been fully developed and researched in recent years. In this paper, the existing fault prediction methods are briefly summarized and studied, and focused on the data-driven prediction method. In the light of the scale of data, all kinds of prediction methods are divided into small sample-based and large sample-based fault prediction methods. The basic principles, advantages and disadvantages of each type of typical analytical methods are summarized, and finally, we put forward the research prospect the development direction of future data-driven fault prediction.

Recently, the modelling of variation propagation in multistage machining processes has drawn significant attention. In most of the recently developed variation propagation models, the dimensional variation is determined through kinematic analysis of the relationships among error sources and dimensional quality of the product, represented by homogeneous transformations of the actual location of a product's features from their nominal locations. In design and manufacturing, however, the dimensional quality is often evaluated using Geometric Dimensioning and Tolerancing (GD&T) standards. The method developed in this paper translates the GD&T characteristic of the features into a homogeneous transformation representation that can be integrated in existing variation propagation models for machining processes. A mathematical representation using homogeneous transformation matrices is developed for position, orientation and form characteristics as defined in ANSI Y14.5; further, a numerical case study is conducted to validate the developed methods.

The complexity of a multi-stage machining process makes it very difficult to assess process plans, from a geometrical point of view. Today, repeated trial runs or very simple simulations are the methods applied to assess a process plan before production release. Due to the cost and time involved in trial runs and the inaccuracy of current simulations, an alternative means should be investigated. This could involve developing accurate 3D simulation methods. This paper proposes a detailed mathematical formulation of tolerance analysis based on searching for the worst case. The performance of two current optimization methods for worst case identification (genetic algorithm and sequential quadratic programming) will be explored through two 3D examples. This paper builds on the model of manufactured part (the MMP) previously proposed by the authors for modeling the different geometrical deviation impacts on the part produced (error stack-up) in a multi-stage machining process. The model cumulates the impacts of various sources of manufacturing errors, hence enabling tolerance analysis. Previously, the same authors also presented a generic formulation for tolerance analysis based on searching for the worst case using the MMP.

The machining quality of a workpiece is determined by many machining process elements, such as the machine tool, fixturing, and process planning, etc. It is especially observed that there is a propagation chain of quality attributes in the multistage machining processes (MMPs). Owing to the propagation chain of quality attributes, the machining quality of the workpiece at a single stage is determined not only by the machining elements of the current stage, but also by the quality attributes of its previous stages. In order to describe the propagation chain, a form-feature-based quality control model for MMPs is proposed. Based on this, a linear error propagation model is established by the use of a rigid-body kinematics method, which can explicitly describe the mapping relationship between machining errors of quality attributes and machining process elements. In order to have some comparison with the error propagation model, the above linear regression model has been built to enable correlated quality attributes to be determined with this error propagation model. A workpiece has been used to show that the error propagation model can effectively describe the propagation chain of the quality attributes.

This paper discusses the in near–real time processing of Global Navigation Satellite System observations at the Department of Theoretical Geodesy at the Slovak University of Technology in Bratislava. Hourly observations from Central Europe are processed with 30 minutes delay to provide tropospheric products. The time series and maps of tropospheric products over Slovakia are published online. Zenith total delay is the most important tropospheric parameter. Its comparison with zenith total delays from IGS and E–GVAP solutions and the validation of estimated zenith total delay error over year 2018 have been made. Zenith total delays are used to improve initial conditions of numerical weather prediction model by the means of the three–dimensional variational analysis at Slovak Hydrometeorological Institute. The impact of assimilation of different observation types into numerical weather prediction model is discussed. The case study was performed to illustrate the impact of zenith total delay assimilation on the precipitation forecast.

Current software defect prediction and code quality assessment methods treat these inherently related tasks independently, failing to leverage their complementary information. Existing graph-based approaches lack the ability to jointly model structural dependencies and quality characteristics, limiting their effectiveness in capturing the complex relationships between defect patterns and code quality indicators. This paper proposes a novel integrated model that simultaneously tackles both objectives using graph neural networks to leverage the inherent graph structure of software systems. Our novelty lies in the first-of-its-kind integration of multi-level graph representations (AST, CFG, DFG) with a dual-branch attention-based GNN architecture for simultaneous defect prediction and quality assessment. Our approach constructs multi-level graph representations by integrating abstract syntax trees, control flow graphs, and data flow graphs, capturing both syntactic and semantic relationships in source code. The proposed dual-branch GNN architecture employs shared representation learning with attention mechanisms and multi-task optimization to exploit complementary information between defect prediction and quality assessment tasks. Comprehensive experiments on six real-world software projects demonstrate significant improvements over traditional methods, achieving F1-scores of 0.811 and AUC values of 0.896 for defect prediction, while showing 9.3% average improvement in code quality assessment accuracy across multiple quality dimensions. The integration strategy proves effective in capturing complex structural dependencies and provides actionable insights for software development teams, establishing a foundation for intelligent software engineering tools that deliver comprehensive code analysis capabilities.

In the last few years, with the development of deep learning theory, researchers have tried to introduce the method of artificial intelligence into the field of software defect prediction (SDP) to improve its prediction effect. To be fed into the neural network, the sample codes are represented as an abstract syntax tree (AST), and the AST is encoded as real numbers. However, in most cross-project defect prediction (CPDP) task, the method for converting the AST into a real number cannot effectively estimate the semantic distance between the ASTs, resulting in a significant reduction in training effects. To solve that problem, we present a new encoding framework, tree-based-embedding (TBE), to convert AST into real vectors and make the semantic gap between the ASTs measurable. To estimate the effect of this encoding method, we promise a tree-based-embedding convolutional neural network with transferable hybrid feature learning (TBCNN-THFL) to perform the CPDP tasks. TBCNN-THFL is fed data encoded with TBE method for learning the transferable joint features between different projects; meanwhile, TBCNN-THFL introduces a transfer component analysis algorithm. Furthermore, the model combines the handcrafted and deep-learning-generated features and then feeds them into the classifier to train a defect prediction model. A sufficient number of experiments demonstrate that TBCNN-THFL is superior to referential models on 72 pairs of CPDP tasks formed by 9 open-source projects.