Gery Fossaert, Principal Simulation Engineer

Simulations allow us to test hypothesis and react quickly to the results, enabling our teams to be agile and efficient as they deliver high quality solutions for our customers where every tenth of a second counts.

Gery Fossaert moved to the UK from Belgium in 2013 to pursue a Master’s degree in Electromechanical Engineering and consequently, a career in motorsport. Gery developed his skills as a project engineer and simulation specialist in the automotive industry at a number of companies, including Jaguar Land Rover and MAHLE Powertrain. After seven years’ experience, he then joined us here at Cosworth, where he has risen to the role of Principal Simulation Engineer.

What drew you to the simulation side of engineering? 

What I really enjoy about simulation is that it gives us an understanding of how our products can be used in real world applications; we can envision exactly what we need to do through a digital medium. When faced with a set of problems to solve, you don’t always understand what the data is telling you at the outset. Simulation allows us to support the analysis of the modelling activity by building a copy as a digital twin. Through this development in the digital space, we can see all of the features the data is showing us, enabling the team to understand what is happening. 

Once there is a good correlation between the digital model and reality, the next step is to find explanations for the results we have achieved. After that, we move onto the final stage: deployment of high-fidelity models in support of the battery development process. We also provide this capability to the wider company, as we can package the models as a stand-alone application so they can run their desired scenarios – for example, if they wanted to assess battery thermal behaviour for high power applications. 

On the technical side, can you tell us about the development path?

Battery modelling can be split into different categories: cell level, module level, and pack level. While these are the different entities, they are also a building block for one another. 

On a cell level, the lowest level building block, we have a combined thermal and electrical model. We can separate these environments and develop them individually within the digital space. We used the test data from the real cell to set corresponding parameters of the digital model which ensures that the model is functioning as accurately as possible.

This stage of development is fundamental to the model at higher levels. When you have a validated initial cell building block, you can then use it at a pack level to produce accurate results. For example, when you get a duty cycle from a customer, you have set parameters to develop around. You can then size the battery correctly to achieve the desired cycle.

Can you tell us about your role, as a Principle Simulation Engineer? 

When I started at Cosworth, the simulation capabilities were tailored to specific projects.  While this structure proved useful, we found there was a need for a tool chain to support the wider business. I brought my previous experience as a model builder, in which I developed models to sell to customers, and I translated those skills into creating tools that were applicable for Cosworth’s needs.

As Principal Simulation Engineer, I put a structure in place to define how we build models according to the standardisation procedures that we use, and I have applied that same technique for all the different battery cell applications. Once we had that in place, we had our building blocks. This enabled us to develop the tool chain to automatically configure packs, reducing the model development time and making the whole process considerably more efficient.

Configuring these packs digitally is done as you would in real life. You take the cell, put it in a module, and use the tools to test it in a virtual environment. Analysing data from the virtual environment forms a big part of our workload, therefore I also develop data analysis tool-suites to streamline how we look at data and get the outputs quicker. An example of this is our battery State-Of-Health (SOH) post-process tool that we apply to the British Touring Car Championship (BTCC) race logs.
 
As part of my role, I lead these activities and act as a mentor to the next generation of simulation and analysis engineers that come through our doors.

When you say that you develop tools, are you developing software or using software to develop tools that run within a program? 

We use MATLAB Simulink to develop our tools. This is a visual software environment where I develop specific tools. The software then compiles and runs the code in the background, allowing for a user-friendly interface within the tools. This method means we avoid having to code everything line by line. In my field, if we had to code it all we would lose clarity and the task would become very convoluted, limiting us to creating even very simple models. 

Using simulation software, with an interface to make visual models, makes it far easier to understand what each component is doing and why you are getting certain outputs, making further development easier and faster. 

What is the benefit of the work you are doing in simulation when it comes to inbound enquiries? 

With the use of high-fidelity models, when a customer with specific duty cycle requirements approaches us and asks how big the pack needs to be, we can give them a very accurate size estimate. You can do this very early on in the process. This also reduces the amount of design iterations you need to go through, which is key to efficient development. 

The models we have created are also applicable to our Battery Management Systems (BMS). We can simulate conditions that allow us to map out running temperatures of a battery, enabling us to know when the BMS needs to step in to restrict the power in order to reduce temperatures. 

Do you see the outcomes of simulation having a big effect on the efficiency of the development process? 

The development of our data analysis tool-suite has been an exploitation of developing analysis tools for simulation models. Deploying these analysis tool-suites on our larger data sets has proven to speed up analysis time and extract Key Performance Indicators from the data – a key example being   the BTCC.

In the BTCC, we gather data from every car across the grid. By the end of the season, we have compiled more than 4000 logs containing all the data captured from the cars – not just related to the battery. To manage this, we’ve developed another program in MATLAB. First the data is captured using the Cosworth’s data acquisition tool from Pi Toolbox. It is then organised and structured by the program to make it easy to navigate. From here, we can develop metrics to analyse the data automatically and pinpoint any potential issues. After each round, we follow our Key Performance Indicators to understand how packs are being deployed in the field and evaluate how they are ageing throughout the seasons. 

What projects have you played a part in so far? 

We currently have on ongoing project, in which we are creating a pack using immersive cooling for Norton Motorcycles. Through this programme, we are further developing our simulation capabilities. We are now applying these tools across all of our current programmes, including the British Touring Car Championship (BTCC) battery and others.

Where do you see battery simulation going at Cosworth in the next few years? 

I hope to see our capabilities develop as we continue to expand our team of highly skilled engineers. We have developed a lot of the backbone of our simulation offering and we are training more people to use it. 

 Will the tool-suite evolve as well?

It will most certainly evolve. Our current portfolio of toolsets mainly comprises of battery simulation models, data analysis toolkits and stand-alone apps. Each of these toolsets keeps on expanding. To further enhance our battery simulation models, we are developing 3D thermo-electric cell models which allows us to study the inhomogeneity of current and temperature with the cell. The data analysis toolkit is shared with our Development and Test team in order to support future cell characterisation and cell ageing programmes. Lastly but not least, our main stand-alone app which forms the basis of our battery sizing toolchain receives updates continuously as we are growing our cell and cooling type model libraries.