Deterministic and stochastic methods in manufacturing process

Manufacturing has changed drastically over the last decades due to adaptation and application of new technologies from different disciplines. The increase of computational power brought artificial intelligence  as counterweight to the often very sophisticated deterministic modelling with new chances and challenges into play and is today with the massive evaluation of data from different sources part of Industrie 4.0.  Collaborating at the forefront of manufacturing technology MSUT Stankin and IWF / ETH Zürich want to mobilize their combined expertise in manufacturing processes to explore the seam between the two different approaches, as it becomes evident that the stochastic methods come to their limits and can be cross-fertilized with physical models.  This allows enhanced prediction and analysis, creation of digital twins throughout the entire manufacturing field, for process steps, process chains, manufacturing equipment, quality assurance etc., T and are the basis for further digitization and utilization of the internet of things.

The goal of proposed project is setting up a joint initiative to explore the most effective and efficient combination for different tasks in manufacturing. Both organizations share their experience, create an inventory of mutual competences and a knowledge pool for best practices. Tools involved are

  1. Artificial Intelligence (AI) methods: Neural Networks, Bayesian optimization, Markov–Decision Processes, iterative learning control, Ontologies, Monte Carlo method etc.
  2. Physical modelling: phenomenological and analytical modelling, FEM and meshfree simulations.

Potential research directions:

  1. Stochastic modelling of spark placement and crater generation in Electrical Discharge Machining (EDM). Using probability function and Fuzzy logic for the influencing parameters like field line concentration, debris, roughness remaining ions in the gap, etc.
  2. Direct Metal Deposition (DMD) process modelling with experimental verification using 10 kW disk laser from Trumpf, available in MSUT "STANKIN". Process controls for extremely high power DMD with powders of different and especially limited quality to minimize the process costs.
  3. Processes with geometrically non defined cutting edges, esp. brittle-hard materials, their stochastic parameter description combined with physical models on the single grain and the macroscopic level taking into account wear and dressing.
  4. Prediction of fractalized supply chains.


Joining expertise would further increase the knowledge through planned tandem work and workshops on models and modelling methods. Through this project it is possible to create a competition which model in the end is best suitable to generate some difficult to obtain results.

The goal of the present PSG is to trigger an active scientific and educational collaboration between academic institutions and help to:

  • strengthen existing collaboration
  • develop applied projects and set up a creative innovative network between students and researchers in both countries;
  • further develop complementary skills and promote academic excellence among graduate, postgraduate and PhD students;
  • support graduate students and PhD candidates promising research through research advisors in both countries;
  • create a basis for future long-term collaboration:

Publications in international scientific and Russian and Swiss industry journals will already result.


ETH Zürich:

Moscow state university of technology "Stankin":