Streamlining multi-scale modeling: a BPMN-based no-code approach for simulation workflows

MUSICODE, an EU research project, aims to create an Open Innovation Platform for OLAE applications. Faced with the challenge of translating theoretical modeling into executable multi-scale modeling workflows, MUSICODE turned to Cardanit for a solution.

By leveraging the BPMN standard and Cardanit's intuitive editing features, MUSICODE streamlined the process, enabling modelers to effortlessly transform MODA workflows into executable simulations, all without writing a single line of code. This innovative approach improves workflow efficiency, fosters greater collaboration, and accelerates progress in the field of materials modeling.

The ContextBringing OLAE innovation with BPMN and multi-scale modeling

MUSICODE is a European research project funded by the European Union’s Horizon 2020 research and innovation program. It’s coordinated by the University of Ioannina and counts 11 partners from both academia and industry, including TinniT Technologies. The goal of MUSICODE is to create an Open Innovation Platform for Materials Modelling in the Organic and Large Area Electronics (OLAE) application domain.

OLAE is a rapidly emerging sector bringing disruptive technological revolutions in energy, electronics, transport, photonics, buildings, lighting, and displays. It also impacts health, wearables, the IoT, and agriculture. OLAE devices are manufactured on flexible polymer web rolls. They represent complex architectures of multilayer stacks of organic, polymer, inorganic, and hybrid nanomaterials.

MUSICODE aims to create multi-scale modeling workflows and address specific problems in organic photovoltaic device development and fabrication. These devices also go by the name of organic solar cells. MUSICODE chose the BPMN standard to design the modeling workflows, and here is where Cardanit comes into play.

The ChallengeFrom MODA workflows to executable workflows

In the field of materials modeling, modeling workflows are usually documented using the so-called MODA workflows. MODA (which stands for MOdelling DAta) is a template for the standardized description of materials models, meant to guide users towards a complete high-level documentation of materials models, starting from the end-user case to the computational details.

A MODA workflow is not executable. So, it needs to be implemented to be simulated. This means either developing the code from scratch to link solvers and data sources together, or facilitating the implementation using an integration framework like the Multi-Physics Integration Framework (also known as MuPIF). Both options require writing some amount of code and having some knowledge about software development. In this context, a no-code approach could make it easier for modelers to go from MODA workflows to executable workflows. This is one of the core ideas of the Open Innovation Platform developed in MUSICODE.

The SolutionMapping MODA workflows to BPMN processes

The MUSICODE no-code approach is based on BPMN since MODA workflow constructs naturally map to BPMN elements. Cardanit has been selected to implement the MUSICODE workflow editor, given its user-friendly editing tools and strong adherence to the BPMN standard.

Thanks to Cardanit, modelers can easily map their MODA workflows to BPMN processes. Tasks within these processes represent physics-based simulation models available in the Open Innovation Platform, while data elements describe the data exchanges between them. Once a BPMN process is complete, the backend service of the Open Innovation Platform can process its XML standard representation to generate the Python code for running the workflow via MuPIF.

The BenefitsNo-code approach for efficient implementation of multi-scale simulation workflows

Modelers using the MUSICODE Open Innovation Platform can implement executable multi-scale simulation workflows without writing a single line of code. This allows them to put all the focus on the interaction between the simulation models. The no-code approach developed with Cardanit can lead to savings in terms of reduced debugging time and reuse of simulation models.

Multi-scale modeling workflows can have a high number of data exchanges between simulation models. This results in managing a high number of Data Inputs, Data Outputs, and Data Associations in the corresponding BPMN process. To correctly convert the BPMN process into Python code, it’s important to set all the bits and pieces of the BPMN data elements correctly. This task, usually daunting and error-prone, becomes quick and easy thanks to the Cardanit Data Flow tool, as the MUSICODE partner TinniT explained.

The figure above shows the multi-scale simulation workflow TinniT created for the Organic Vapour Phase Deposition (OVPD) fabrication of OLEDs. The workflow involved connecting five tasks in series and configuring them with the appropriate simulation models. Using the Data Flow tool, TinniT quickly set up the required input items to match the Data Input task requirements. They sorted the data elements, placing process parameters, which were the same for all simulation runs, graphically on the left-hand side of the workflow. Each of these items was linked to all five tasks. Then, they organized the parameters to be varied for the study on the right-hand side. Subsequently, they assigned a Data Output (simulation report file) to each task. Finally, after designing the simulation workflow and generating the corresponding MuPIF Python code, TinniT started the simulation study in the MUSICODE Open Innovation Platform by adjusting parameter values.

TinniT Technologies

SME TinniT, founded by Battelle Institute research assistants in Frankfurt/Main, Germany, is now based in Karlsruhe, specializing in prototyping, development, technical calculations, and software solutions. Its core expertise lies in simulating heat and mass transfer through cellular, phase-change, and biological materials. TinniT also develops multiscale materials through 3D printing and coating with micro-porous materials. In research, they focus on improving their CFD solution methods (TinFlow) and other numerical tools.