Advancements in advanced modelling of complex products and systems

Bart H.M. Gerritsen (Fac. Industrial Design Engineering, Delft University of Technology, Delft, The Netherlands)
Imre Horvath (Fac. Industrial Design Engineering, Delft University of Technology, Delft, The Netherlands)

Engineering Computations

ISSN: 0264-4401

Article publication date: 2 March 2015



Gerritsen, B.H.M. and Horvath, I. (2015), "Advancements in advanced modelling of complex products and systems", Engineering Computations, Vol. 32 No. 1.



Emerald Group Publishing Limited

Advancements in advanced modelling of complex products and systems

Article Type: Editorial From: Engineering Computations: International Journal for Computer-Aided Engineering and Software, Volume 32, Issue 1


Our intention with compiling this Special Issue for Engineering Computations was to cast light on the advancements in terms of the methods and tools for modelling products and systems. It is well known from the literature that there are many new research problems and open issues concerning complex product and system modelling dueto architectural and operational complexities, as well as to methodological deficiencies and limitations. They can prevent providing a purposeful and effective support for advanced engineering activities. For instance, the development of testable concepts or design variants may be blocked; insightful structural or behavioural analysis may be obstructed; or the product manufacturing and real-life operations cannot be simulated or even realized. Though some of the mentioned challenges may not belong to the front line of scientific research, they need to be resolved in order to support the current engineering practice.

This underpins the theme selection for this Special Issue. Achieving progress in advanced modelling of complex products and systems is step-by-step process. With regards to the above-mentioned challenges, we must also take into account that not the state of technology (including tools and techniques) is the most severe limiting factor, but the state of our knowledge and understanding. It has been recognized that some of the traditional methodologies and methods are not sufficiently powerful to address complex problems. For instance, they can only be applied to a specific set of design problems, or to a limited set of enabling technologies. If the used methodologies can any more not fulfil the needs under the influence of emerging and rapidly evolving product development technologies, then we are forced to reconsider our methodologies.

The specific aim of our editorial work was to investigate what concrete methodological deficiencies and mismatches had been identified by design researchers and practitioners, and to explore new approaches, which are more adequate to complex product and systems. Our first impression has been that the issue of modelling complex products and systems need more advanced methodologies. It must also be mentioned that methodological limitations may appear not only in the design stage of product development, but also in other downstream stages, such as engineering simulation, manufacturing process planning, product realization, supply, and delivery. Whenever disruptive new technologies emerge, their exploitation may be slowed down by a sluggish (lately starting and intuition driven) development of matching methodologies, tools, and techniques.

Structuring the Special Issue

To arrive at a reasonable arrangement of papers, we imposed a thematic structuring in this Special Issue. We identified two groups of topics and sorted the paper accordingly.

The first group includes papers that address issues associated with the morphological (geometric, topological, and structural) attributes of the design object. In this context, on the one hand, the main issue is the lack of proper methodologies and methods for modelling and analysis of products, and the absence of adequate tools and techniques, and on the other hand, the growing need for advanced morphological and functional descriptions of products.

The second group comprises papers that discuss issues associated with the application context of product models. They deal with methodological challenges that typically present in the application-dependent downstream activities of product realization, including complex product models of engineering analysis and simulation, manufacturing process planning, and supply chain design. The papers also deal with the challenges of digital representation of products, product data, encoding properties, and deal with the reuse of digital product data and the handling of the product data during its life cycle.

It has to be mentioned that the above grouping should not be viewed as a general classification, or taxonomy. Based on a broader collection of papers, different classes could be formed. In our case, there was no one single criterion to make the two groups categorically distinct or mutually exclusive, for the reason that certain issues of morphological description and design may occur also in the downstream activities. In other words, it may well be that a topic considered in the first group may alsio be considered in the second group based on slightly different sorting criteria, and vice versa.

Contribution of the papers

The first paper in the first group, entitled “Sensitivity analysis in optimized parametric curve fitting”, has been submitted by Oscar Ruiz et al. (2015). They elaborate on the topic of human-supervised, high-quality fitting of open uniform B-spline to some noise data set, with the consideration of computational accuracy and feature correctness. Typically, some distance functions are developed to find the optimal fit. The work presented by Ruiz et al. proposes to investigate the sensitivity of the goodness-of-fit for some candidate parameters in order to find the best parameter to control the fitting process. The enhanced sensitivity analysis completed by the authors resulted in a very efficient optimization process, while it avoids having peaks, curls, etc., so that an additional curvature analysis can be disregarded. In the sensitivity analysis, an optimized number of control point turned out to be critical. Ruiz et al. adapted existing methods and provided their results based on these.

The second paper, contributed by Jean-Philippe Pernot et al. (2015), is entitled “Thin part identification for CAD model classification”. They discuss the comparison and selection of thin parts in CAD models that require special attention during finite element model generation and analysis. A central problem in the work of Pernot et al. is taking geometric data from a CAD system as input and performing shape and feature recognition. Their work was done with the objective of re-using carefully optimized meshes for thin parts. Towards this end, the authors had to tackle a search problem: given a set of search criteria, identify all fitting models (parts) in a database. As a first step, the authors developed a classification scheme to help identify candidate models (parts). They applied scale-less mesh criteria to find the fitting meshes in a database. This work points at the need for additional mesh optimization knowledge that could be re-used in the problem at hand. Based on the evaluation of this optimization knowledge, the re-useable optimal database model can be found. Operating on STL files, a first prototype has been demonstrated, and incorporation in CAD or PDM tools is anticipated.

The third paper discusses a somewhat resembling research topic, i.e., the phenomenon of sliver-like finite elements. This paper was submitted by Ruding Lou et al. (2015) under the title “Filleting sharp edges of multi-partitioned volume finite element meshes”. The proposed method associates semantic information with volumetric meshes. The aim was to support the identification and removal of potential problem elements from the mesh using the associated additional information. The authors suggested alternate elements to prevent potential problems. They also intended to circumvent the need to return to the CAD model, which would increase model-analysis round-trip dramatically. To achieve their objectives, they applied filleting to modify potential problematic mesh elements, assuming that high stresses can appear on inadequately shaped elements. A mesh operator that does this can be integrated into regular meshing tools. This paper demonstrates that greedy solutions may contribute to the solutions of an underlying fundamental problem at methodological level. The framework they propose for meshing can be integrated with several CAE, FEA, and PLM tools.

The last paper in this group, entitled “Functional restructuring of CAD models for FEA purposes, was contributed by Ahmad Shahwan et al. (2015). The authors studied the augmentation of CAD data with additional functional information to support the transformation of CAD data for finite elements-based structural analysis. More specifically, they studied semantic enrichment of finite element meshes and problem formulations. They studied the functional (kinematic) meaning of problems, and their solutions, in order to use functional knowledge to drive problem solving. Their argumentation is that solutions can be obtained more quickly, if engineers understand the functional background of a component. They used digital mock-ups to define the context for their functional inference.

The first paper in the second group, reporting on the work of Roberto Raffaeli et al. (2015), is about “Virtual prototyping in the design process of optimized mould gating system for a specific form of die casting”. The authors look for an advanced modelling and simulation method and tool to study and optimize material in-flows and through-flows in high pressure casting. They tried to get to insights in how casting materials fill up in the case of complex die casting. The motivation comes from the fact that filling up often induces problems in the case of complex alloy products. In the lack of proper understanding of material flows, these problems should typically be solved by trial-and-error approaches, which is a spill of many resources.

The second paper in this group, entitled “Numerical analysis of geometrical and aerodynamic enhancements of a birdlike wing”, starts out of a real-life design challenge (Harik et al., 2015). Co-authored by Ramy Harik et al., this paper discusses an interesting study into the modelling of airplane wings with bird-like features such as flapping, and into morphing wings by applying remiges (“feathers”) and bending mechanisms in the wing. Any attempt to model such (combined) features challenges the existing model formulations for conventional wings. In this work we see a clear demonstration of the earlier described need to modify or replace a methodology, before the actual work can be fruitful. For practical reasons, the authors focused on both take-off and landing situations. Nevertheless, the treatment of the non-conventional features shows immediate modelling difficulties. Reductions and limitations (for instance on the winglets) were needed to manage the complexity of the application. Following a greedy approach, the authors foresee the opportunity of tackling the winglet problem in a further study, for which yet again, extension or replacement of the current methodologies will be needed.

In their structural analysis, design engineers primarily focus on anisotropic and orthotropic behaviour of these materials, and do not consider how individual fibres distribute in the material matrix at micro-scale, like material engineers do. To abridge, new micro-to-macro scale models are essential. The last paper placed in this group, presents the work of Huang et al. (2015) under the title “Generalized periodic surface model and its application in designing fibrous porous media”. It discusses the development of a general mathematical model to create complex structures from fibrous porous media. Such a model can be used to model, simulate (e.g. by using Matlab), and understand the complex properties in different directions at various scales. The model proposed by Huang et al. is an instance of a periodic surface model, with emphasis of the role of (woven meshes of) fibres from a 3D physical phenomena point of view. Their objective is to compute joint mesh-based force fields and computational fluid dynamics-based flow patterns. A limitation yet to overcome is the application of parameters for compressed fibres, as a result of mechanical deformation in the joint model.

Reflections and takeaways

What these papers have in common is the need to overcome methodology mismatches first, before finding a solution to the actual problem. The main takeaway we could identify is that a review of methodologies, methods, and tools is needed in order to be able to go beyond routine modelling. an evaluation of process planning procedures is recommended if any change is introduced in the product modelling methodology, methods, or tools. Several papers in this Special Issue deal with meshing and the connection between CAD-modelling and finite element analysis. Apparently, a huge amount of knowledge goes into the creation of finite element analysis meshes that it pays off to consider re-usage of the meshed parts. This knowledge capturing and knowledge re-use interconnects the design context with the application-dependent context.

In this regard, we can identify another takeaway: it is advantageous to capture and document the mesh optimization knowledge for re-use. If possible, prepare the whole mesh for re-use. Appropriate mesh database searching and evaluation procedures are needed to support re-use; not only for a mesh as a whole, but also for individual regions within a mesh. Towards the re-use of optimally meshed CAD-model parts, the work of Shahwan et al. could elegantly be combined with the work of Pernot et al. Whereas automated intelligent de-featuring is being foreseen by Pernot et al. for FEA purposes, Shahwan proposes to add functional features for the sake of semantic enrichment of the design intent, to guide engineers in their interpretation.

Once a complex product has been designed and its structure has been verified and validated, process planning and manufacturing should be able to accommodate the concerned implementation as-is. Changing the product just because the manufacturing process is not properly understood and controlled is not only a spill of resources, but also destroys validated product design knowledge and obstructs re-use. This leads us to a next takeaway: Seeking opportunities for methodological enhancement in the context of the whole life cycle of a product is becoming an important issue.

In designing and engineering of complex products and systems, various disciplines work concurrently on shared tasks. Their collaboration can be supported by model-based work. Physical, augmented, or virtual prototypes are produced as containers of multi-disciplinary and cross-disciplinary engineering knowledge that can be exchanged and shared. Physical and virtual prototype-based simulations should provide cross-disciplinary assessment of correctness. It can be assumed that an evolution affecting one discipline can stimulate a snowball effect throughout all the disciplines involved. We believe the papers included in this Special Issue have created a possibility for this. We are grateful to each of our authors for their contribution. We appreciate their efforts and nice collaboration. We are also indebted to our reviewers who helped us to achieve a high-quality Special Issue.

Bart H.M. Gerritsen and Professor Imre Horvath, Faculty of Industrial Design Engineering, Delft University of Technology, The Netherlands


Harik, R., Nicolas, A., Dassouki, M. and Bernard, A. (2015), “Numerical analysis of geometrical and aerodynamic enhancements of a birdlike wing”, Engineering Computations, Vol. 32 No. 1, pp. 86-101

Huang, W., Didari, S., Wang, Y. and Harris, T.A.L. (2015), “Generalized periodic surface model and its application in designing fibrous porous media”, Engineering Computations, Vol. 32 No. 1, pp. 7-36

Lou, R., Pernot, J.-P., Giannini, F., Veron, P. and Falcidieno, B. (2015), “Filleting sharp edges of multi-partitioned volume finite element meshes”, Engineering Computations, Vol. 32 No. 1, pp. 129-154

Pernot, J.-P., Giannini, F. and Petton, C. (2015), “Thin part identification for CAD model classification”, Engineering Computations, Vol. 32 No. 1, pp. 62-85

Raffaeli, R., Favi, C. and Mandorli, F. (2015), “Virtual prototyping in the design process of optimized mould gating system for high pressure die casting”, Engineering Computations, Vol. 32 No. 1, pp. 102-128

Ruiz, O.E., Cortes, C., Acosta, D.A. and Aristizabal, M. (2015), “Sensitivity analysis in optimized parametric curve fitting”, Engineering Computations, Vol. 32 No. 1, pp. 37-60

Shawan, A., Léon, J.-C., Foucault, G. and Fine, L. (2015), “Functional restructuring of cad models for fea purposes”, Engineering Computations, Vol. 32 No. 1, pp. 155-176

Related articles