#12 Innovative business models for IoTwins PaaS in manufacturing

The test-bed aims to validate innovative business models that bring resources, available in the cloud (e.g. (High performance computing, Digital Twins, AI networks, etc..), accessible to applications, related to the machine monitoring business, running on manufacturing  machines. This testbed wants to demonstrate that innovative technologies such as Deep Neural Networks (DNN) and cloud-based access to HPC resources can open new opportunities to more pervasive and transversal business models. That is to be achieved through techniques that
1. speed up the training of a monitoring system (in its capability of recognizing and anticipating patterns of process degradation) and
2. facilitate the access to the above services from edge twins, also to decouple implementation from IoT device-specific characteristics.
The proposed business model is based on a SaaS Cloud solution that the customer can use on a “pay-per- use” basis. All the data read from the manufacturing machine and commands sent from involved edge nodes are collected and stored on the cloud. The collected data are used to train a DNN to obtain a predictive model of the degradation of the monitored process.
Training is accelerated by the usage of the cloud-accessible IoTwins PaaS infrastructure, which can exploit also HPC resources. Once a reliable degradation model is available, the trained IoTwins digital twin can be used to test the behaviour and effectiveness of several control algorithms. Once the test has identified the best control algorithm, it can be compiled, on request, by using the specific toolchain needed to generate the executable code for the specific destination edge hardware platform.
Thus, different edge vendors can take advantage of the features and training acceleration made available by the use of our cloud-based infrastructure.

  1. Demonstrate that the controlled process degradation can be modelled by a digital twin based on a DNN

  2. Validate a “pay-per-use” solution to manufacturing line owners

  3. Define rules and interfaces between edge and PaaS models that allow the usage of the PaaS from different edge nodes and platforms

  4. Speed up the training of monitoring systems in their capability of recognizing and anticipating patterns of process degradation

  5. Facilitate the access to the above mentioned functionality and services from the edge

Involved Partners

#11 CAMP NOU replicability towards smaller scale sport facilities
#1 Wind turbine predictive maintenance