Contact Twinn

CONTACTER Twinn

As a leading automotive manufacturer Ford is fully aware how improvements and efficiencies in their operations can be sought and achieved. One department in Ford that is particularly successful in this regard is PTME (Power Train Manufacturing Engineering) where simulation has been key to successful predictive analysis that has already delivered real gains in productivity; this is set to continue into the future.

Simulation in the 1980s – When Ford Met WITNESS

John Ladbrook, Ford’s simulation technical expert, has been responsible for overseeing the implementation of the PTME simulation strategy since the early 1980s. This role also included sponsorship of many engineering projects across various universities and resulted in John’s appointment as Fellow with Cranfield University in 2001 and Honorary Professor at the University of East London in 2016.

The emergence of simulation back in the eighties allowed engineers to influence the planning process in a way that was previously rarely considered. Simulation models made transparent the effects of change and investment and allowed a wide variety of options to be tested cost effectively. Experimentation and thorough investigation of cause and effect led to a greater body of evidence that could be presented to senior management to justify decisions. 

“Sharing and effective communication of the simulation results was, and still is, absolutely essential to establishing successful modelling,” John argues.

This approach of using predictive simulation models was so successful that Ford soon mandated that investment in simulation would be a pre-requisite for all future program verification – and so the PTME invested further in Lanner’s commercially available WITNESS software to support model development and satisfy Ford’s demand for predictive analysis.

This commitment to upfront analysis in advance of any production or manufacturing decisions led to regular meeting forums based around the simulated results. Productivity soared, but each major study project took approximately 600 hours to complete – effectively 4 months of full-time work.

This led to a situation where the demand for simulation was far greater than the supply. Ford needed new ways to break the deadlock and John was tasked with solving this problem without significantly increasing the business headcount.

New Challenges

As part of their program to continually improve the efficiency of simulation and reduce costs, Ford opted to develop its own productivity-enhancing software, known as the Ford Interactive Replacement Simulation Tool (FIRST). FIRST was built to not only decrease the time required to answer engineering questions, but also to enable simulation capability at a more accessible skill level. Key to the design was a simple spreadsheet interface that an engineer, without deep modelling and simulation skills, could populate with data and then rapidly run experiments which, in turn, fed back a set of easy to understand results.

“Our FIRST system was used across all projects for new and existing machine production lines,” John recalls, “By this stage, Ford’s simulation capabilities were widely recognised as being world class and were seen as a beacon for the industry. At this stage, only two full time engineers were responsible for Ford’s simulation strategy, supplemented by various university located projects, making such accolades all the more impressive!”

Sponsoring local universities created a mutually beneficial relationship between Ford, Academia and the students. This allowed Ford to gain access to resources and expertise while also providing students with real world context to incorporate into their thesis and experience in a world-class automotive manufacturing engineering team. Since 2007, Ford’s PTME simulation team has grown from 2 people to 13, five of whom had previously been students working on projects at Ford.

This collaborative approach remains in place today – enabling chosen students to gain a greater knowledge of cause and effect through the use and manipulation of simulation and in some cases leading to long-term employment within Ford.

Further Refinement

Ford’s development of simulation stepped up a gear when, in 2007, an additional tool was developed to allow for greater flexibility and even easier access and learning for new users. This model became known as the Ford Assembly Simulation Tool (FAST). It refined the lessons learned with FIRST and utilised more of the open connectivity power of WITNESS software. The result was a consolidated, more flexible tool that could be used more widely across powertrain assembly and engine machining production lines. Management were demanding an improved alignment to reality and the improvement through FAST meant this demand could now be met.

2479-2_Dagenham_engine_line.jpg

Ford's Dagenham Engine Line

The simulation of each project study with FAST was now taking under three months to complete. This efficiency was a great improvement and simulation was consequently becoming more and more in demand. In response, a greater number of WITNESS nodes were deployed across the Ford networks, allowing the team to answer more questions with wider experimentation that improved results.

Over the next few years, the development of FAST continued; utilising HPC (High Performance Computing) clusters in which banks of powerful processors provided John with the computing power he needed to further increase experimentation rates and team productivity. Before this could be achieved, development of WITNESS and the Ford infrastructure had to be made and this was successfully achieved through collaboration between Ford’s simulation team, Ford’s IT department in North America and Lanner’s development team. The HPC experimentation shaved another 2-4 weeks off the delivery time for each project study, but again the key outcome was delivering more analysis and decision support for management. In the first full year of using HPC, Ford’s simulation team experimented across a total of approx. 100,000 hours of WITNESS simulations. A figure that has since quadrupled by the end of 2016!

Another factor crucial to the team’s success has been that of delivering easy to understand analysis in a format that others could easily interpret and relate to. Investing in the development of sophisticated tables, charts and analytical calculations, Ford’s simulation experiments now seamlessly generate engineering, manufacturing, execution and performance measurement reports in the exact format that Ford management understand and are familiar with.

This has had the effect of further encouraging senior management to ask even more business questions of the simulations; success breeding success! Fast configuration of scenarios, fast experimentation and fast analysis and creation of reports underpins the success of Ford’s FAST simulation capability.

Recent Work – a Symbiotic Simulation

In 2015, the PTME in conjunction with the University of East London PhD began developing a “symbiotic solution” with an aim to shrink the required configuration time for each study even further. With the abundance of connected data sources available, it becomes feasible to tap into live manufacturing systems and condition the data directly into the formats consumable by the simulation.

Ford leveraged the API interface to WITNESS to connect and control models using interfaces from other systems. The result are models configured and driven directly by live and accurate data sources effectively generating a digital twin of real world systems.

Models will increasingly be fed in this way in the era of smart manufacturing as modelling becomes embedded in operational as well as strategic decision making. The symbiotic simulation of the real facility will take the latest data and plans and constantly assess future performance.

Is Simulation Maturity in Sight?

Today, John states, “Software underpinned by WITNESS predictive simulation software and a talented team of ten have allowed the PTME to cover more ground and to develop a bigger and more detailed picture of actions and consequences than ever before. Advanced modelling as well as simulation question-and-answers now work hand-in-hand, allowing significant cost-savings. It is Ford management’s thirst for simulation that has enabled and allowed these advances.”
“Success is also marked by the growing use of simulation in the plants and, most of all, in other departmental meetings where we frequently hear ‘what does the simulation say?” John remarks. “In my opinion, we are by no means at the peak of the maturity curve – having spent the best part of my career establishing a world class simulation capability at Ford, I am excited about how future developments will enhance our achievements further.”

Looking into the future, the goal is to continue to reduce the elapsed time required to complete a study project. If analysis can be returned in less than one business day an even wider range of questions can be asked and answered in enough time to influence decisions. To do this PTME aim put more intelligence into the upfront assimilation of the data. Simulated outputs will continue in the same formats but the tools used to present results will change; the team are already looking at alternative options for analytics, both locally and deployed across networks.

Increasing Demand

It is also forecasted that as business complexity and demand for volume increases steadily year-on-year, the use of predictive simulation will increase along with the scope of business questions. John’s team are looking to introduce more advanced analytics capabilities into Ford’s current tools to be able to handle their increasing complexity.

Redesigning the user interface will allow PTME to guide production engineers, who have typically been internal customers of the team, to use the simulation tools. Removing the additional complexity in the project process means that the correct implementation decision can be taken every time. And this simplicity will be essential to the next milestone in Ford’s simulation program.

In the near future Ford are aiming to allow every plant to run their own simulations on demand. Using FAST, on-site engineers can expect to be able to look back 3 months, get statistically stable data, condition the data for the experiment, run the simulation and obtain results. They should then be able to use AI and manual intelligence to guide their actions based on their modelled predictions.

Key Ingredients to Ford’s Simulation Success

John identifies a number of key factors in making simulation a success at Ford:

  • Establishing the right methodology and developing a streamlined process to enable fast simulation along with efficient data entry.
  • Modelling at the correct level of detail. The level of detail needs to be determined by the questions that are to be asked of a model. It is easy to waste time with too much detail. John states “My advice to others looking at building their simulation capabilities is to consider how to model robustly. Do not get drawn into emulation (modelling every detail) which is time consuming in terms of building the model, gathering input data, and which may yield little return in terms of model accuracy relative to the input.”
  • Integrate simulation modellers with the plant teams. The simulation team has now been co-located with manufacturing, so that communication can become more direct – and that engineers on both sides can feel integral to the overall process.
  • Develop reporting systems that communicate well. The right report will be understood immediately by the recipient and enable quick actions.
  • Develop methods to run model experiments quickly. The development of an HPC scaled deployment with Lanner has allowed for fast results and the solution is scalable.

Simulation at Ford – the Bottom Line

John considers that the use of simulation is key to driving efficiency at Ford.

“Have we achieved everything we set out to achieve? Absolutely, and more. I am immensely proud of the simulation capability we now have, and the fact that it has driven multi-million-pound savings in production programs across the world over the years is testament to the software capability and our methods of use.”


Loading blog comments..

Post a Comment

Thank you, your comment is awating approval
Submit