© TU Delft

Multidisciplinary design optimisation MDO dashboard

Multidisciplinary design optimisation (MDO) identifies optimal design configurations that simultaneously incorporate the requirements of multiple disciplines, by optimally trading-off the competing needs of diverse factors and drivers, such as structural design, material usage, floor layout, installations and so forth. The principles of MDO are in contrast with traditional design optimisation which usually addresses a small number of objectives and a single design aspect, meaning that the design team do not gain a clear picture of the impact of optimisation on other design aspects.

The aim of this research was to develop a user-focused MDO framework to optimise a parametrically-defined problem, such as the design of a building, addressing multiple disciplines together. It can be applied with any design and analysis software, for example structural, energy and light analyses. The MDO framework consists of three parts:

– a database containing the parameters defining different design alternatives

– a search engine that allows the user to search through the design alternatives

– a web-based user interface (the dashboard) that allows the parties in the design process to evaluate the alternatives.

Via the dashboard, design team members and clients can obtain a clear overview of the data from the design optimisation process, enabling them to see how each parameter is influencing the optimised design.

 

The user interface is a critical element to get right: the dashboard was highly effective in allowing the user to search visually through large data sets (typically many thousands of options) and to see relations between parameters that could otherwise be overlooked.

The MDO framework can be easily scaled up by dynamically adding machines to its computation cluster, subject to any contractual constraints regarding the jurisdiction in which project data can be stored.

The MDO is a decision-supporting tool: it does not remove the need for engineers to undertake the initial design, transform it to a parametric model, and verify the results of the MDO.

The complex nature of building design can be characterised as multi-parameter, multi-discipline and multi-objective. Multidisciplinary design optimisation (MDO) allows a design team to incorporate all relevant disciplines simultaneously, and to consider the interaction between those disciplines. It finds an optimum for all disciplines in parallel together, in contrast with the traditional approach of combining together separately optimised designs from each discipline. The MDO approach leads to a better global optimum of the problem, although it introduces significant non-linear complexity to the solution, including many local optima.

The MDO framework presents enhanced capabilities for multidisciplinary projects including the ability to incorporate collaborative optimisation to determine the optimal option for the project client and relevant stakeholders.

Tools based on the principles of multidisciplinary design optimisation (MDO) can strengthen and support collaboration between different technical specialists.

Taking advantage of advances in computing and storage that are now mainstream, MDO can allow more efficient delivery of optimised designs to clients and facilitates better communication with clients about the factors influencing the design of a building.