Dimensional accuracy improvement of FDM square cross-section parts using artificial neural networks and an optimization algorithm

Fused deposition modelling (FDM) is the most extended additive manufacturing technique up to date. In FDM, a thermoplastic material is extruded through a nozzle to form layers, and the final geometry is the result of consecutive superimposed layers. However, it is difficult to obtain an adequate dimensional accuracy for some applications due to the characteristics of the process. This paper proposes a method for increasing accuracy of the distance between parallel faces on FDM manufactured prismatic parts, consisting in replacing the theoretical values of CAD model dimensions by new optimized values. For this purpose, a model has been developed for predicting the dimensions of the manufactured parts, based on design characteristics. Particularly, this work has used an artificial neural network combined with an optimization algorithm, to determine the optimal dimensional values for the CAD model. Subsequently, CAD model is redesigned according to the dimensions provided by the optimization algorithm, and the part is manufactured. The results show that the application of this methodology allows for a reduction in manufacturing error of approximately 50 % for external dimensions and 30 % for internal dimensions.     

Optimization of fused deposition modeling process parameters for dimensional accuracy using I-optimality criterion

Fused deposition modeling (FDM) is one of the widely used additive manufacturing technologies because it has flexibility and ability to build complex parts. The accuracy of parts fabricated by FDM is greatly influenced by various process parameters. FDM process has a complex mechanism in building parts and often poses difficulty in understanding adequately how conflicting FDM parameters will determine part quality and accuracy. Sectors such as medical implant, telecommunication, electronics and aerospace require increasingly higher levels of dimensional accuracy. Thus, traditional methods of ensuring quality do not effectively address global markets and customer’s needs. This study proposes an I-optimality criterion for the optimization of FDM process parameters in order to address the limitations of the commonly used traditional designs. This study also aims to develop mathematical models in order to establish nonlinear relationship between process parameters and dimensional accuracy. The results show that I-optimality criterion is very promising technique in FDM process parameter optimization. Confirmation experiments show that the proposed method has great advantages in the aspect of both accuracy and efficiency compared with traditional methods proposed in previous studies.     

Guidelines for building 2D CAD models of turned components in CAD-CAPP integration

EXCAP is a knowledge-based process-planning system for rotational components. CADEXCAP is software which has been developed to integrate EXCAP with CAD systems via IGES (Initial Graphics Exchange Specification). CADEXCAP deciphers a full 2D model of a turned component developed on a CAD system and then via IGES it fully automatically generates a product model in a format suitable for EXCAP. This paper describes the guidelines which have to be followed at the CAD modelling stage of the components which are to be used by CADEXCAP. Following these guidelines will guarantee the automatic extraction of complete product models. This not only includes the geometry but also includes nongeometric information such as toleranced dimensions, geometrical tolerances, manufacturing instructions and surface finish.     

Surface roughness prediction in fused deposition modelling by neural networks

Fused deposition modelling is a proven technology for the fabrication of both aesthetic and functional prototypes. The obtainable roughness is the most limiting aspect for its application. The prediction of surface quality is an essential tool for the diffusion of this technology, in fact at product development stage, it allows to comply with design specifications and in process planning it is useful to determine manufacturing strategies. The existing models are not robust enough in predicting roughness parameters for all deposition angles, in particular for near horizontal walls. The aim of this work is to determine roughness parameters models reliable over the entire part surface. This purpose is pursued using a feed-forward neural network to fit experimental data. By the utilisation of an evaluation function, the best solution has been found. This has been obtained using a feed-forward neural network for fitting the experimental data. The best solution has been founded by using an evaluation function that we constructed. The validation proved the robustness of the model found to process parameters’ variation and its applicability to different FDM machines and materials.     

A method for assessing geometrical errors in layered manufacturing. Part 1: Error interaction and transfer mechanisms

Geometric accuracy of components is one of the most important quality characteristics in layered manufacturing processes on which most rapid prototyping (RP) techniques are based. Layered manufacturing is an approximate fabricating process in which the final geometric error of the physical part is affected, not only by the approximation technique used, but also by the fabrication process. Errors that occur in one layer could propagate and transfer to other layers causing an accumulated error effect in the process. In this paper, a concept of disturbance error is introduced to describe the effect of accumulated errors in the fabrication process. A physical model is presented to describe error interactions and error transfer mechanisms in the layered manufacturing process. A geometrical model is developed using surface approximation techniques to describe the relationships of the geometrical errors. It is shown that although the complexity of the part geometry is not directly related to the manufacturing process, it will affect the geometrical errors of the part produced.     

Extraction of manufacturing information from design-by-feature solid model through feature recognition

Feature recognition is the key to the computer-aided design (CAD) and computer-aided manufacturing (CAM) integration to build a computer-integrated manufacturing system. There are two approaches to CAD feature recognition: platform-dependent and platform-independent. In the platform-independent approach, the part’s geometrical data are extracted from a neutral file such as DXF, IGES, or STEP. In contrast, the platform-dependent approach extracts the information of the design features directly from a design-by-feature solid model through the object-oriented model of a part. This paper explains a platform-dependent approach which is implemented to translate design features into manufacturing information. This approach begins with simplification using the suppression of fillets, and clustering non-intersecting design features is done. Then, the rule-based method is employed in order to recognize machining features. Finally, the needed manufacturing information such as tool accessing direction, dimensions, material removal regions, and geometrical and topological data is recognized. The application of the proposed system would be exhibited in generating machine path code for rapid prototyping and CNC machines and providing a database for computer-aided process planning. The proposed system was implemented on Autodesk Inventor and successfully tested for many complex 3D models.     

Repeatability of linear and radial dimension of ABS replicas fabricated by fused deposition modelling and chemical vapor smoothing process: A case study

In this research work, an effort has been made to study the influence of fused deposition modelling (FDM) and chemical vapor smoothing (CVS) process parameters on the selected linear and radial dimension as well as on repeatability of acrylonitrile butadiene styrene (ABS) replicas as a case study. The study highlights that orientation of parts on FDM build platform, part density (part interior style) and interaction between these two parameters significantly affect the accuracy of selected dimensions. Shrinkage has been observed in the selected radial dimension of the prototypes, but there is a positive deviation in the linear dimension from the desired value. The CVS process reduces both the dimensions slightly due to reflow of the material. Optimum parameter settings that were different for both linear and radial dimensions have been investigated using Taguchi’s L18 orthogonal array. The IT grades of ABS replicas prepared by this combined process were found to be consistent with the permissible range of tolerance grades as per ISO standard UNI EN 20286-I (1995) and DIN 16901 for plastic materials. Finally, optimum level of process parameters that simultaneously minimizes the deviation in both the dimensions have been found out using response optimization module of Minitab 17 software and the results obtained have been verified by performing the confirmation experiments.     

熔融沉积制造高能效工艺参数优化方法

为降低熔融沉积制造过程中能量消耗,提高能量利用率,针对熔融沉积制造参数优化问题进行了研究。建立了熔融沉积制造过程能量效率函数,以喷嘴温度、成型层高、空走速度和打印速度为优化变量,以最大能量效率为优化目标,利用田口法设计正交实验,得出能量效率较高的工艺参数组合。在此基础上,利用BP神经网络建立能效优化模型,采用自适应小生境遗传算法对模型求解。研究结果表明,要实现熔融沉积的高能效制造,需在较高的喷嘴温度和成型层高的情况下,采用较低的空走速度和打印速度。     

Machining strategy development and parameter selection in 5-axis milling based on process simulations

Fused deposition modeling (FDM) is one of the most popular additive manufacturing technologies for fabricating prototypes with complex geometry and different materials. However, current commercial FDM machines have the limitations in process reliability and product quality. In order to overcome these limitations and increase the levels of machine intelligence and automation, machine conditions need to be monitored more closely as in closed-loop control systems. In this study, a new method for in situ monitoring of FDM machine conditions is proposed, where acoustic emission (AE) technique is applied. The proposed method allows for the identification of both normal and abnormal states of the machine conditions. The time-domain features of AE hits are used as the indicators. Support vector machines with the radial basis function kernel are applied for state identification. Experimental results show that this new method can potentially serve as a non-intrusive diagnostic and prognostic tool for FDM machine maintenance and process control.     

Part decomposing algorithm based on STL solid model used in shape deposition manufacturing process

Shape deposition manufacturing is an innovative rapid prototyping technology combining CNC machining and deposition processes and in which if a surface of the part is visible from the build direction it can be directly shaped into desired geometry on part material; if the part features are undercut or face away from the build direction, these part features will be sculptured into the support material in order to form a negative cavity at first, then the part material is deposited into this negative cavity in order to replicate these features into part material. This process is cyclically repeated until all segments have been built. Finally, the sacrificial material is removed, and the finished part is left. This innovative technology has the benefits of commercial rapid prototyping technology; meanwhile, it inherits the accuracy and high surface quality of traditional machining process. For this process to be realized, it is essential to decompose the CAD solid model into a series of manufacturable segments which have no undercut surfaces at the data pretreatment step. In this paper, a solid decomposing strategy is proposed with a detailed discussion on the classification of the segments based on the value of triangle face normal vector, and then an algorithm for auto-decomposing a CAD solid model in STL data format is developed.     

Polishing mechanism for ABS parts fabricated by additive manufacturing

Fused deposition modeling (FDM) is a process of fabricating three-dimensional physical models by layered manufacturing. However, the surface quality of acrylonitrile butadiene styrene (ABS) built with FDM technologies is not acceptable and is not satisfactory for most general engineering purposes. In this study, a polishing system was demonstrated to enhance surface finish of ABS parts fabricated by FDM. Polishing mechanism for ABS parts was investigated. The features of this system include high polishing efficiency, no waste chemicals, ease of operation, with excellent dimensional accuracy, and low equipment costs. Improving surface roughness of ABS parts by the use of acetone vapor not only has required dimensional accuracy but also has high process stability. A great reduction in average surface roughness for ABS parts about 98% was obtained. The polishing mechanisms for ABS parts fabricated by additive manufacturing were investigated in this study.     

Computer-aided geometrical dimensioning and tolerancing for process-operation planning and quality control

This paper describes an extension of a model which determines an optimum set of dimensions and tolerances for machining processes at minimum manufacturing cost. This optimisation minimizes the cost of scrap, which is a function of manufacturing tolerances, as the objective function. Requirements of design sizes, geometrical tolerances (including both form and position) and machining allowances are expressed mathematically as constraints for the optimisation. A computerised trace method has been extended to determine the relationships between geometrical tolerances and associated relevant manufacturing dimensions and tolerances. In addition to the manufacturing cost, the model takes into account manufacturing sequence, distribution of manufacturing dimensions, process capabilities, tolerances, design sizes, geometrical tolerances, machining allowances and optimum scrap level. The resulting computerized interactive system can be used not only in process planning, but also in quality control.     

Tolerance evolutionary model and algorithm in product growth design

To guarantee the successful transformation from product functional requirement to geometry constraints and finally to dimension constraints between components in a product, an evolutionary tolerance design strategy is proposed on the basis of automation technology in product structure design. In the first part of this paper, the theory of the growth design and the process of tolerance evolutionary design are introduced. Following the evolution of product structure, product tolerance grows from its initial state, defined by accuracy requirements, to its final state, defined as dimension tolerance and geometric tolerance. In this growing process, the basic units in the product growth design, known as functional surfaces and their nominal features, are used as evolutionary carriers. With the help of these basic units, the method for the construction of a two-layer correlation network is proposed. In the second part, the tolerance assertions to assist tolerance evolutionary design are given, based on which the basic process for an evolutionary design of dimensions chain and geometric tolerance are presented. In order to optimize the allocation of the dimension tolerance, a mathematical model is developed in which a correlated sensitivity function between the cost and the tolerance is created. In the model, the design cost, the manufacturing cost, the usage cost, and the depreciation cost of the product are used as constraints to the tolerance allocation. Considering these costs, a multifactor cost function to express quality loss of the product is developed and is applied into the model. The minimum cost is used as the objective function, and the depreciation cost in the objective function is expressed by the discount rate—terminology in economics. The aim was to achieve a final and ideal balance around assembly, manufacturing, and usage through the control of product precision. In the last part, the successful usage of the proposed tolerance evolutionary strategy in the incremental growth product design is demonstrated through a design example.     

Determination of fabrication direction to minimize post-machining in FDM by prediction of non-linear roughness characteristics

Rapid Prototyping (RP) is a technology that generates physical objects directly from CAD data. As a prototype can be rapidly built by using this technology, the product cycle is remarkably reduced in manufacturing. However, the surface quality of RP processed parts is not sufficient for general mechanical parts because of the use of the layered manufacturing process. Therefore, additional post-machining processes are required such as grinding and coating. But, these processes are time-consuming and detrimental to the original part’s geometry. In this paper, a methodology for determining fabrication direction which aims to minimize the post-machining process in FDM (Fused Deposition Modeling) is proposed. In order to enhance the accuracy and the computational speed in determining the fabrication direction, a surface roughness model that is considered even non-linear surface roughness characteristics is presented. And, a required post-machining volume is newly introduced as a objective function. Also, a genetic algorithm is applied to obtain a reliable solution. Finally, it is verified that the proposed approach can be very efficient with the developed system.     

自顶向下设计及快速成形技术应用

首先,以Pro/E软件为设计工具,采用自顶向下的现代产品设计理念,使用骨架法对鼠标模型进行数字化设计及装配,从而实现鼠标外观造型和结构的一体化设计;然后,将鼠标各个零部件的三维模型在Pro/E中软件保存为STL格式;最后,使用三维立体打印设备,采用熔融沉积制造工艺(FDM),对鼠标各个零部件的模型进行快速成形,自动、快速、直接和精确地将计算机中鼠标零部件的三维实体设计转化为实物模型。从实物模型中可对所研发的产品外观造型和结构进行改进和优化,提高产品设计和研发效率,缩短设计研发周期,从而节约设计和制造成本。     

Enhancing moulds manufacturing by means of reverse engineering

Modern CAD/CAM techniques together with five-axis high-speed milling allow to reduce moulds manufacturing time and costs. Nevertheless, in order to put a mould into use, operations of manual finishing and fitting are still always required. Such operations, performed manually by mould makers, modify the milled surfaces of moulds. Reverse engineering techniques can be employed in quality control to evaluate dimensions and geometrical tolerances on moulds after machining and fitting. Changes in the product’s shape are sometimes decided after a mould has already been machined. In such cases, if possible, the mould maker directly modifies the mould. Thus the final real geometry of the mould does not reflect the one of the original CAD model. The aim of this paper is to point out the benefits of non-contact quality control and to illustrate a procedure, based on reverse engineering techniques, to reconstruct and update the mathematical model of the mould after it has been polished and fit. The procedure was tested on a mould for the production of a plastic camera body that was previously inspected by means of a structured light scanner.     

CADEXCAP: Integration of 2D CAD models of turned components with CAPP

EXCAP is a knowledge-based process-planning system for turned components. This paper describes a software package called CADEXCAP, which has been developed to integrate EXCAP and CAD systems via IGES. CADEXCAP deciphers a full 2 D model of a turned component developed on a CAD system and then, via IGES, fully automatically generates a product model in the EXCAP model format. It is clever enough to extract not only the geometrical information but also the other required information, such as geometrical and dimensional tolerances, surface finish and manufacturing instructions.     

Fused deposition modelling (FDM) process parameter prediction and optimization using group method for data handling (GMDH) and differential evolution (DE)

This paper presents the research done to determine the functional relationship between process parameters and tensile strength for the fused deposition modelling (FDM) process using the group method for data modelling for prediction purposes. An initial test was carried out to determine whether part orientation and raster angle variations affect the tensile strength. It was found that both process parameters affect tensile strength response. Further experimentations were carried out in which the process parameters considered were part orientation, raster angle, raster width and air gap. The process parameters and the experimental results were submitted to the group method of data handling (GMDH), resulting in predicted output, in which the predicted output values were found to correlate very closely with the measured values. Using differential evolution (DE), optimal process parameters have been found to achieve good strength simultaneously for the response. The mathematical model of the response of the tensile strength with respect to the process parameters comprising part orientation, raster angle, raster width and air gap has been developed based on GMDH, and it has been found that the functionality of the additive manufacturing part produced is improved by optimizing the process parameters. The results obtained are very promising, and hence, the approach presented in this paper has practical application for the design and manufacture of parts using additive manufacturing technologies.     

On Line Model Accuracy Inspection of Model Maker Rapid Prototyping Using Vision Technology

The principle of rapid prototyping (RP, or layer manufacturing) is to fabricate a 3D object layer by layer. Vision technology has been employed and has demonstrated its ability to measure the single-layer profile dimensions of the Model Maker (MM) RP system model. The on-line profile dimensional inspection and accuracy analysis of multiple layers is presented in this paper. The image-processing algorithms, developed to measure the layer profile dimensions, were adapted to process the captured profiled image of multiple layers, layer by layer. Hence, the layer profile dimensions of multiple layers were inspected on-line and layer by layer. The problems of vision inspection associated with colour mixing of multiple layers are addressed. A colour CCD camera was employed to resolve the colour mixture problem, but it is too expensive and too much memory is required. A new image difference algorithm was proposed to solve the problem of colour reflected from adjacent layers. The results demonstrate the promise of on-line profile dimensions inspection and accuracy analysis using vision technology.