Design Automation Lab

A new car in the manufacturing area.
Volvo Cars is one of  LiU:s partners

Product development was revolutionized by the introduction of the first CAD software in 1963, a digital drawing board called Sketchpad. The premise of the innovation was to automate a manual, repetitive and erroneous draft-drawing process.

In hindsight, the inventor of Sketchpad, Ivan Sutherland maximized the technological possibilities available in the 1960s. changing industry as it became vastly easier to draft designs and this spurred major innovations over the next half-century. The premise has changed since the 1960s. There is no longer a struggle in draft design, but many new fields need an innovative touch.

Two of the current industry challenges Design Automation Laboratory study are:

  • Engineering-To-Order (ETO) products and
  • Complex Engineering Products

In order to face the above challenges, we focus currently on following research areas:

  • Knowledge Based Engineering
  • Multidisciplinary Design Optimization
  • Machine Learning
  • System of Systems Engineering


Research notices

Industry challenges 

Two of the current industry challenges Design Automation Laboratory study are Engineering-To-Order (ETO) products and Complex Engineering Products.

Engineer to Order

The conflicting demands of increased customization of products and decreased costs and delivery times poses challenges in product development and manufacturing companies. Many companies attempt to manage the complexity of customization with platform- and product family design where the physical attribute of the product is broken down into modules to be shared between product variants. Still, product developing, and manufacturing companies are decelerated by repetitive work, misinterpretations, assumptions and uncoordinated processes which hinders the achievement of mass customization where increased customization does not imply higher costs and lead times. Especially for customized products, a significant amount of engineering resources is demanded for each customer order - a main challenge to overcome in mass customization. Another challenge and well-established problem within product development are the lack of knowledge in the early stages of the product development process, where the design freedom is at highest, and vice versa in the later stages, often referred to as the design paradox. Research within product development and the development of computer aided technologies often has this paradox as its vantage point and aims at increasing the knowledge in the early stages in order to produce better products more efficiently.

The design paradoxThe design paradox

Complex Engineering Products

CAD software have been incrementally improved to contain much more than just a digital draft board. CAD is today at the centre of product development where a digital representation of the physical product can be generated early in the product development process. Various characteristics of the product can be computed and analysed by utilizing physics-based tools such as FEM, CFD, dynamic system simulation, etc. Each needing various information contained in CAD. The connection between various fields is itself a major drawback in terms of repetition and extensive lead times. 
For instance, during the design phase the processes of geometry definition in CAD and other disciplines are filled with repetitive operations. One example of this is that CAD geometries are needed for computer aided engineering (CAE). However, the geometries needed for CAE analysis need to be simplified, meshed, pre-processed, analysed and then post-processed, where each stage is typically managed by different individuals in the organization. Furthermore, the process is highly iterative and repetitive, and in any give stage there might be a need to iterate back one or several steps, see Figure 1. Applying various types of digital automation techniques can be of great value to speed up the response time of the experts operating at various stages.

. Example of a highly repetitive part of the CAD-CAE integration. Example of a highly repetitive part of the CAD-CAE integration.


Knowledge Based Engineering

Knowledge Based Engineering (KBE) is a popular method applied to enable design automation of repetitive CAD operations. KBE defines a wide range of methods and processes. We define KBE as:

Automating non-creative design tasks by utilizing object-oriented programming.

At Design Automation Laboratory we have developed a new way to efficiently automate design with geometry primitives called High Level CAD templates (HLCt), namely:

  • Generic CAD model can be placed and automatically adapted to various geometrical shapes without needing to manipulate any parameter.
  • The method is generic and not specifically designed for 1 or 2 specific CAD tools.
  • Designers start from a blank sheet and configure component by component, like a virtual LEGO system. Add new LEGO bricks to your library whenever you want and thus your design options are unlimited.


Multidisciplinary Design Optimization

Multidisciplinary optimization (MDO) is defined by Giesing et al. (1998), as

A method for designing complex technical systems and subsystems that consistently exploits synergies between interrelated phenomena

A classic example with two domains is the strength and aerodynamics of an aircraft wing which, for reasons of strength, should be rough, but slim due to aerodynamics. In practice, however, more disciplines must be taken into account and also more compromises must be handled. MDO today has a relatively small distribution in Swedish industry and is often not really multidisciplinary.

MDO has many advantages, but a practical problem is that MDO is a numerical process with many computationally demanding iterations. To address this, effective methods are being developed within the research group.

Since different subsystems generally have conflicting optimal solutions, a holistic perspective is necessary to achieve an optimal design from a holistic perspective. In addition, there are always several and conflicting requirements for the product, which is why the so-called Multi-objective optimization (MOO) is an attractive way to tackle the problem, as even a large number of so-called Pareto optimal solutions are generated. One way of dealing with conflicting requirements and a changing set of requirements from customers is to simultaneously adopt a development strategy based on modularization and product families.


Machine Learning

There are currently many active research fields exploring various ways to streamline, automate, and rationalize the Product Development Process (PDP). Amongst these fields are Knowledge-Based Engineering (KBE) methods in order to achieve Design Automation (DA) by enabling the reuse of knowledge, ultimately focusing on automation of “repetitive, non-creative design tasks”. KBE-driven DA is based on rules created to describe the possible iterative and repetitive scenarios. KBE is thus suited for well-defined product ranges where the design space can be approximated beforehand based on customer requirements. However, for highly customized products there is a limit as to how far KBE can be stretched and determine the product range with pre-defined rules.

Machine Learning (ML) has seen a significant increase in interest over the last two decades, academically as well as amongst industries and consumers, due to its notable presence in modern technology in forms such as computer vision, speech recognition and recommendation systems and industrial applications such as robotics. With the rising prominence of ML, an increased interest in its application in the fields of engineering design and product development is inevitable.

Systems of Systems Engineering

Complex engineering products are connected to their operational environment as well as with each other, and nowadays, it is often observed that most real-world applications are based on a synergetic approach of several interconnected systems. What is more, complex products can have long development and production times, and thus, a further challenge is to be able to account for uncertainties like the introduction of a new technology or future market changes. Lastly, the envisioned role of a complex product may easily change within its life cycle, and hence, the new operating conditions can lead to an underperforming or even unusable system. By looking a yet-to-be-designed product as part of System of Systems (SoS), the development team can identify critical design factors early in the process where there is still time to make decisions, while overall, designers can also ensure a better performance and a customized product according to a set of customer-specific needs.

Compared to the traditional Systems Engineering (SE) approach where the requirements are fixed, System of Systems Engineering (SoSE) is based on exploring a much larger design space and it is expected to lead to solutions that can provide better resilience to constantly changing requirements. More specifically, the consideration of the system’s participation in a larger ensemble, allows analysts to identify several SoS constellations that can deliver the desired capabilities, and therefore, this makes it possible to relax or even remove some of the initial system requirements. Accordingly, an SoSE approach enables manufacturers to align a product with their business strategy, product portfolio, technology, and competence development plans, while in general, it can also help them adjust the same product for the needs of more than one customer. 


Download zip-file

2D drawings

This tutorial includes instructions and data for supervised learning av 2D drawings.

Download zip 2D drawings

WebGL configurator

This tutorial contains instruction and data for a simple WebGL configurator.

Download zip WebGL configurator.


Customized Products

These tutorials are used in the Master Course Design Automation for Customized Products.

Download zip Customized Products.

Partners and financier

Some of our partners and financier

Siemens Turbomachinery
Hitachi ABB
Volvo Cars

Fordonsstrategisk forskning och innovation
Processindustriell IT och Automation
Produktion 2030


Researchers in the Laboratory