Optimal rubber injection moulds through cloud-based simulation
Rubber parts are present in a multitude of components used in a wide variety of industrial sectors of which transportation is the most relevant one. The market size in Europe for rubber parts amounts to approx. € 18 billion per year.
The objective of this application experiment is to improve the quality of the parts aiming at zero-defect (i.e. scotch, flash, weld lines and air-trapped) design while minimising injection time. The methodology applied for achieving this goal is based on a lean definition of the app to be running on the cloud. This means that user interaction is kept to a minimum, no visual interface is needed and reporting is done by automatically created reports.
Technically, it is expected to reduce the complexity of the rubber injection design process to allow rubber component manufacturers to optimise mould designs and accelerate ramp-up processes through user-friendly access to computational cloud services. Economically, the time and cost savings due to a simulation-based process are estimated to amount to 10 to 15 percent of the profit margin.
Currently, the end-user starts with a 3D CAD model of the rubber part to be produced. Based on their experience they design a preliminary mould and its heating system. Then, based on few optimization loops (typically from zero to five loops) the mould design and operation parameters are optimized. In this step, specialised CFD commercial software (imported from the plastic manufacturing sector) is used. After that, the mould is built and a start-up phase is needed to optimize the operation parameters by trial and error using the manufacturing machine. Typical lengths of these phases are: 1 month for mould and process design and optimization, 2-3 months for mould manufacturing and 1 month for start-up. Typical cost distribution is: 50 percent raw material, 5 percent energy consumption, 15 percent manufacturing equipment amortization and 30 percent personnel.
As described above, the current process already involves different simulations varying parameters such as mass flow rate, heat flux distributions, etc. However, due to the number of parameters and their value ranges, a full simulation of all parameter combinations, requires HPC resources, which are not available to the end user today. Thus, the mould design, mould manufacture and ramp-up of the manufacturing take 4 – 5 months, not guaranteeing zero defects with the first moulds manufactured.
The approach of this Application Experiment is to run a design-of-experiments suite of simulations on an HPC cluster, derive a meta-model from it that can then be used and evaluated for different parameter combinations guiding the engineers to much better starting positions, and ultimately allow for achieving zero-defect manufacturing with the first mould produced. This promises a 25 percent time reduction for the design and optimization cycle, as well as for the start-up phase and quality inspection. The latter time reduction is because one side-result is the area where a defect can be found. Therefore, visual quality inspection can be focused in analysing these areas in detail and not the whole part.
The two end users in this Application Experiment see the main technical impact in the process innovation: from a trial-and-error approach based on specialists’ experience to an explicit knowledge-based approach using meta-models that capture deep insight into the rubber injection process and optimised parameters. This leads to an improvement of final product quality due to the minimization of part defects. A better quality means reduction of inspection time, reduction of start-up time and reduction of rejected parts.
For the ISV the technical impact is on making their tools available via a Cloud-based solution, easily accessible via the CloudFlow Portal. This enables ITAINNOVA to address all members of the Spanish cluster of rubber industries (ASICE with more than 80 companies) via a unified cloud solution.
Economically, the novel approach translates into an improvement from 10 percent to 15 percent of the profit margin obtained for each mould. Half of this margin is attributed to time reduction during start-up and optimization stages (i.e. less energy, less workforce, less rubber) and the other half to the reduction of part rejections (i.e. less wasted rubber, less time for quality inspection). More importantly, it is expected to produce almost no defect parts anymore, thus also reducing scrap. Plus, productivity will increase considerably not only due to shortened design and start-up phases but also due to minimized filling time – thus more parts can be produced in the same amount of time with higher quality. The improved competitiveness is expected to show increasing sales figures and to create new jobs (expected 5 percent in three years, which means one new job for a SME like MIJU).
ITAINNOVA as ISV will increase in the number of projects for existing and new customers through the offer of the new service and the development of new workflows in different industrial fields. For example, since the workflow is also valid for non-Newtonian materials injection (i.e. plastics or die casting in addition to rubber) to overall market that can be addressed with this solution amounts to € 180 billion per year in Europe, ten times bigger than the rubber market alone. This will contribute to additional revenue of € 150,000 per year for the next five years and the creation of three new jobs at ITAINNOVA. The HPC resources needed for carrying out the simulations will be provided by UNIZAR-BIFI with a revenue forecast of € 40,000 per year. This will allow the creation of two new jobs in the next three years.