Advanced Computing in the Age of AI | Thursday, March 28, 2024

Paving the Way to Simulation Democratization 

<img style="float: left;" src="http://media2.hpcwire.com/dmr/democracy.png" alt="" width="95" height="72" border="0" />As globalization continues to dominate the business world, companies need to develop new and innovative products to stay ahead of the competition and remain successful in a fast-paced market. To meet these demands, leading companies have turned to engineering simulation to accelerate and optimize development processes and manage increasing product complexity...

As globalization continues to dominate the business world, companies need to develop new and innovative products to stay ahead of the competition and remain successful in a fast-paced market. To meet these demands, leading companies have turned to engineering simulation to accelerate and optimize development processes and manage increasing product complexity. This demand for advanced technology has prompted the need for highly-qualified computer-aided engineering (CAE) staff, so hiring and retention is a main priority.

In smaller organizations where the simulation team may consist of a few specialists or even one expert losing one employee to a career with, for example a larger company, might be devastating if vital engineering knowledge and workflow processes have not been captured properly.

A rapidly changing world

Because business activities have become more global, competition today can come from across the street, from the other side of the globe, from a multinational and established leader or even from a start-up. Such fierce competition calls for a growing differentiation between products that in turn increases products complexity. To succeed in this highly competitive world, most companies have adopted and deployed engineering simulation tools to accelerate and optimize product development processes. This quickly increases the demand for highly educated CAE staff.

To address these four business drivers (globalization, fiercer competition, increasing product complexity and lack of qualified resources), innovative companies are employing two key business initiatives:

  • Develop a lean product development process to reduce costs and react quickly to any competitive move or emerging demand.

  • Boost engineering productivity to compensate for the lack of existing or affordable resources. This results in doing more product development with existing, or possibly even less, resources

Today, competition is not only at the product level, it extends to hiring and retaining people with the necessary CAE skills.

Agile companies adjust processes

Developing a lean and iterative product development process while boosting the engineering productivity of a company is very challenging with a traditional testing approach. However, the need to streamline product development provides an ideal opportunity to more widely and systematically deploy engineering simulation and requires that enough CAE-educated resources are available to small and medium-sized enterprises (SME). The solution is to increase the productivity of existing software while reducing the systematic need for highly trained and specialized resources to use them. This can be referred to as customization and automation.

  • Customization is the adjustment of general-purpose tools to fit the specificities of an industry, a company and/or an application. It includes adding particular features to set up the model such as using specific laws to capture nonstandard material properties and operating conditions as input, and output specific quantities which are more relevant and meaningful for the company. Customization boosts software productivity by making the tool directly usable and specific for more applications and by obtaining immediately exploitable results.

  • Automation is the simplification of the simulation process by minimizing the number of operations necessary to set up a simulation and, in the ideal case, by complete suppression of any user interaction. Automation opens the door to large-scale use of engineering simulation accessible to a much wider audience of engineers and designers. “Remote” or “web-based” automation is an extension of this concept that allows anyone in a distributed company to use an automated environment from wherever they are located.

Customization and remote automation of existing software solutions capture knowledgeable analysts’ expertise and enable more engineers to take advantage of simulation to boost their engineering productivity. It provides more flexibility to the product development process. This democratization of simulation is perceived by innovative companies that have widely embraced engineering simulation as the avenue to more quickly achieve some key business initiatives. CIMdata reached similar conclusions in early 2012 - the acceleration of technology trends allows numerous small and medium size companies to amplify the benefits of democratization. A much wider range of organizations can profit from democratization.

Next--Democratization is a three-stage evolution-->

Democratization is a three-stage evolution

Companies that democratize simulation employ the following sequential but partially concurrent steps.

  1. Customize the solution to extend and simplify its adoption

General purpose software that is efficient, accurate and reliable typically uses a language familiar to CAE-educated analysts that is common across industries. However, this language is less explicit for production engineers and SME managers familiar with industry-specific concepts. Many tools don't focus on specific capabilities crucial for some industries but that are irrelevant for most others. To address this challenge software vendors propose ways to customize their tools, but these solutions are usually complex and often necessitate advanced programming skills. To be truly impactful and effective, general purpose codes should offer an environment that enables analysts to easily adjust the tools in terms of geometrical entities management, materials modeling, operating conditions and post-processing. Customizing results to industry-specific needs and languages is necessary to facilitate and accelerate the communication and interpretation of simulation output and can save precious time for the analyst.

  1. Integrate necessary solutions in a single environment

Customizing a solution is no longer enough as solutions become increasingly complex. Today’s products achieve their challenging objectives and provide companies with a sustainable competitive advantage by making use of different physics (solid, fluid, electromagnetic, acoustic, etc.) often at different levels (nano, micro and macro). In addition, because of mergers and acquisitions or simply fast expansion, even small and medium-size enterprises regularly involve different locations where software might be used.

The multiplication of software solutions is a major obstacle to the democratization of simulation as it massively complicates the exchange of data between codes as well as software license management. Customized environments must not be limited to large organizations or a few expert users if democratization's full impact is to be realized. Customization and automation should be carefully integrated in a common environment to address issues of multiple physics and scale so that organizations of any size can benefit.

  1. Automate the process

A systematic automation of any part of the modeling process is key as it immediately increases the productivity of experts who do not need to repeat the same routine tasks. In addition, automation minimizes the risk of mistakes due to the tedious repetition of these tasks. Automation captures the company best practices and facilitates sharing these practices among company experts while opening the door for non-CAE-educated staff to obtain controlled access to simulation. However, some training is crucial to ensure proper interpretation of simulation results.

 

A recent Aberdeen Group Research report illustrates that best-in-class companies compile best practices through customized workflows

Customization, integration and automation require supporting technologies to achieve democratization of simulation. In the past, attempts at democratization rarely met initial expectations because of technical and technological limitations.

Recent technology enables democratization

The rapid evolution of some technologies in the last few years has greatly reduced obstacles to democratization.

  • Affordable and remotely accessible, (quasi-) unlimited computational power dramatically decreases the computational bottleneck and relieves the constraint that restricted simulation to a few analysts.

  • Streamlined automatic meshing techniques considerably reduce the entry barriers for non-expert users and accelerate the workflow to obtain accurate and reliable results.

  • Development of web-based applications amplifies collaboration through efficient simulation data management (SDM) and PLM integration. These applications facilitate the exchange of information, and customized or automated environments, within a company or between companies. They also simplify the management of centralized or shared software licenses that help rationalize the use of tools.

  • Emergence of easy remote post-processing on smart phones and tablets unleashes the use of simulation to travel outside of the computer room and go into the field, production facilities, manufacturing centers, or maintenance and testing rooms so that simulation can be employed in previously untapped activities.

  • Common programming languages such a Python greatly accelerate the learning curve so that an engineering force can customize and automate the working environment. The availability of numerous libraries on the web facilitates and extends this adoption.

These major technological advances are boosting the demand for modeling power and simulation results while making simulation accessible to a new group of users.

Next--A Measurable Success-->

A measurable success

A democratization initiative that requires intense customization and automation must have clear and accepted ways to measure its success. Different key performance indicators (KPI) are used by companies to fully grasp the success or failure of this activity. Typical KPI includes:

  • The number of projects an analyst can be involved with during a year. As customization and automation lead to delegation of parts of the simulation work, democratization will not be successful if a large fraction of the analysts’ time cannot be allocated to other projects that have not yet benefited from simulation. Through customization and automation, an analyst is typically expected to multiply by three the number of projects he is working on.

  • The number of projects benefiting from simulation early in the product development process dramatically increases. The increase comes both from analysts engaged with new projects and from designers gaining some experience with simulation and using it for projects not important enough to involve analysts. This combined approach is leading to the systematic use of simulation early in the design process. The number of projects that will benefit from simulation is typically multiplied by a factor 4 to 6.

 

Through integration and customization, complete virtual prototyping for packaging can be achieved within a day, leading to systematic product optimization and reducing both cost and environmental impact.

  • The number of people actively and effectively involved with simulation is the ultimate measurement of the democratization process as long as all those engaged in simulation are getting concrete and measurable results from the investment. Leading companies are trying to increase their engineering force engaged with simulation from 0.5 to 5percent.

  • The innovation rate greatly benefits from a large-scale deployment of simulation by a systematic use of simulation early in the design process. Innovation is amplified as confidence is gained from a robust design optimization approach before the release of any new product. The evolution is an intangible asset difficult to measure but nevertheless remains a clear expectation of companies travelling this path.

  • A decrease in warranty costs is one of the long term results of large-scale adoption of simulation, among others through a democratization of the simulation process. Companies have reported a decrease in warranty cost by a factor 4 after deploying simulation.

Democratization of simulation boosts small and medium-size entreprises' engineering productivity

The combination of increasing product complexity and a lack of simulation-educated resources necessitates more software customization and process automation. Rather than employing the few, skilled analysts exclusively for their simulation skills, SMEs can use experts to customize, document and automate engineering workflows so they can be delegated to a much larger group of engineers and designers. Some companies have reported that up to 75 percent of their analysts work is delegated to non-CAE-educated staff through proper customization and automation.

When engineering knowledge is properly captured and delegated to non-expert users, the highly trained analysts can be redirected to other more lucrative investigations. In the worst case scenario, losing these skilled employees but capturing their knowledge decreases corporate risk.  

EnterpriseAI