Supercomputers for manufacturing getting it right

by Julian Fielden (pictured), MD of OCF plc

Today, western manufacturing organisations are facing increasing pressure on many levels - the rising cost of oil, environmental concerns, soaring food prices and continual increases in global competition from Eastern Europe, India and China.   

The confidence of manufacturing firms is now so severely dented 18,000 manufacturing jobs will be lost in the second quarter of the year according to the CBI.  

As part of wider restructure and business development plans, manufacturing organisations are wisely investing in new technology now to support their operations through the tough times ahead.   

As a result, the manufacturing industry has become one of the strongest growth sectors for supercomputers.   

Supercomputers most commonly built using a simple low-cost interconnecting cluster of COTS (commodity off the shelf based PCs, servers or workstations) are becoming more prevalent in manufacturing organisations to store, process, analyse and visualise ever-increasing amounts of data from the production process.   

Manufacturing organisations can use this supercomputer driven data analysis in a range of areas:

• Internally, to more quickly identify opportunities to operate more efficiently, cut costs, increase staff and machine productivity, increase output and improve workplace health & safety, etc 
• For customers, to reduce materials usage, reduce testing stages of production and ultimately reduce time-to-market for goods 
• To support this newly found insight, manufacturing organisations are also turning to supercomputers to power a number of 3D and 4D simulation techniques. 
• 3D simulations (such as CAD) enable manufacturing organisations to build computer aided 3D models of products cars, boats, buildings, consumer goods, etc.  Bringing data to life, it enables manufacturing organisations to more easily identify trends, problems and opportunities.  

By testing those models against a range manufacturing criteria such as time, the models become 4D and enable an advanced level of Insight.     

By example, OCF customer, The University of Birmingham, launched a supercomputer last year that enables it to provide automotive and aerospace organisations like Rolls-Royce with valuable research data (derived through 3D simulations) on welding, casting, heat treatment and forming techniques more quickly.   

The University is also providing organisations with valuable research data on the performance of manufacturing materials under stress more quickly for example stretch testing, cracking and general movement.   

More specific types of simulation techniques include: 

Digital Human Modelling:

Digital Human Modelling helps manufacturing organisations simulate human behaviour in relation to their product designs or manufacturing operations (in the hope of catching problems early on in the design cycle when it is less costly to make changes).   

For example, companies in industries such as automotive and aerospace are using DHM to design assembly operations which reduce worker related fatigue and stress injuries related to repetitive tasks. 

Computational Fluid Dynamics (CFD)

Supercomputer-powered CFD simulates the interaction of fluids and gases with the complex surfaces of buildings and vehicles. 

Lola Cars, the UKs largest and most successful racing car constructor is using the power of an OCF-built supercomputer to help design its future race cars.    The supercomputer enables Lola Cars to use CFD software to simulate airflow over a racing car in a virtual wind tunnel environment.  This enables the aerodynamic performance of a racing car to be calculated as well as allowing new prototype shapes and configurations to be tested quickly.   

Another OCF customer, Nexia Solutions, the organisation preparing to launch the new National Nuclear Laboratory (NNL) for the UK, is also using CFD to create more detailed and accurate models to support its customers work the Likes of Sellafield Ltd (formerly British Nuclear Group) and the Nuclear Decommissioning Authority. 

In this instance, CFD enables safer, more efficient and cost effective completion of essential nuclear projects.   

Finite Element Analysis (FEA)

FEA is a computer simulation technique used in multiple industries, such as aeronautical, biomechanical and automotive, for engineering analysis.  

A common use of FEA is for the determination of stresses and displacements in mechanical objects and systems. Such analysis allows the determination of effects such as deformations, strains, and stresses which are caused by applied structural loads such as force, pressure and gravity. 

FEA enables entire designs to be constructed, refined, and optimised before the design is manufactured improving safety, reducing costs and potentially increasing revenue. 

Next Steps

With rising pressure on manufacturing organisations it is important to get a supercomputer implementation right first time.  Therefore, they must carefully consider selecting a qualified, experienced and knowledgeable integrator to manage the design, technology and management of that supercomputer.   

Technology

Manufacturing organisations working with their chosen integrator - must decide on the most appropriate technology: 

• Unless proprietary solutions offer overwhelming advantage in terms of performance and it is clear that vendor support will be available over the life of a project, integrators should be designing supercomputers based upon open technology which can be developed without proprietary lock-in going forward 
• It is common sense to build supercomputers using Blade Servers which use up to 50 per cent less floor space and up to 58 per cent less energy than traditional servers  
• Manufacturing organisations should demand integrators include software which will help drive efficiencies and performance from the supercomputer.  The software should take full responsibility for scheduling, managing, monitoring and reporting of cluster workloads  
• Compliance regulations are forcing many companies to store data for up to thirty years.  With that in mind, integrators must build scalable storage solutions.  Manufacturing organisations must also ensure integrators introduce a scalable, high performance file system with Information Lifecycle Management (ILM) capabilities, which enables additional storage capacity and performance to be added and operational in minutes with no interruption to users or applications

Management

Finally, manufacturing organisations must ensure their chosen integrator can provide post implementation management and support - OCF for example - offers a Cluster Management and Support Service which enables customers to focus all available IT department resources to non-cluster related queries and user problems.   

Conclusion

With no end in sight to the pressure facing manufacturing organisations, supercomputers have found themselves a new home.   

Manufacturing organisations are increasingly harnessing supercomputers to address potential internal inefficiencies, to develop products at a reduced cost but at increased quality. 

However, organisations must tread carefully to ensure success involve the right integrator, select the right hardware and ensure the right management and maintenance support is in place.