Construction of integrated particle rendering environment for large scale data visualization

Technical simulations are rapidly becoming a fundamental requirement in virtually all sectors, from medicine to energy, aerospace and beyond. In these, there is a growing requirement for real-time imaging and the need to simulate the design, process or even the behaviour of materials under certain circumstances in order to determine how a finished product, device or system may operate.

The driving force behind such visualisations is the need to accurately visualise an irregular conglomeration in highly detailed finite element (FE) models, so as to begin to understand the complexities of not only those themselves, but also how they may interact with the world around them. This might include behaviour under fault conditions, as well as operation in the normal state, and any point in between these two, and opens the way for particle-based rendering to become a major problem-solving technique. Professor Koji Koyamada is at the forefront of this exciting field.

Awareness of the limitations of particle modelling techniques for computational fluid dynamics simulation and visualisation drove Koyamada to investigate volume rendering and the many applications that it can be used in. The main issue with volume rendering is the sheer amount of information that requires computation, and this meant new techniques were required as computer power was set to a large extent. With increasing interest in the modelling of complex datasets – including advanced fluid dynamics – Koyamada became interested in computational fluid dynamics (CFD) as a means to accurately determine the flow characteristics of a gas or liquid in motion. This application has enormous potential in the modelling of not only fluid, but massed-particle systems too, but in order to produce meaningful results, smooth particle hydrodynamic techniques (SPH) were employed. However, it was found that although SPH could be used to create a simulation of complex behaviour, it results in a huge array of particles, which hinders the simulation process. Koyamada employed particle-based volume rendering (PBVR) for visualisation of the huge particle data.