Skip to content

RESULTS

project RESULTS

The concept of the RUBY project is to provide a tool that combines, in an integrated and harmonized framework, advanced techniques for monitoring, diagnosis, prognosis, control and mitigation for FCSs related to stationary power generation. Such tool will be made ready for implementation into industrial processes and for market deployment, embedding general features which are relevant for both SOFC and PEMFC technologies.

TESTING
MODELLING
IMPLEMENTATION

Design of Experiment

Experimental testing

EIS/PRBS-based monitoring

Fault diagnosis

Lifetime estimation

Optimal control

Mitigation strategies

Hardware modification

On-board implementation

MDPC scheme with eis function

Ballard Backup System
Sunfire µ-CHP System

Main schemes of Ballard Backup System (left) and Sunfire μ-CHP System (right)
with the EIS
perturbation (p) and control functions (f ). 

EIS box prototype

hardware upgrade

Former embedded board were replaced by TS-7680 industrial board and mechanical allocation was changed to guarantee integration. Analog-Front-End board (AFE) memory capability was doubled to have 20k samples of voltage and current measurements. A new PCB was design in order to realize on the same board the actual AFE and DC-DC converter (24V – 5V) for AFE power supply; a parallel analysis was done for a simplified AFE version  to reduce dimension for better integration.

TESTING ACTIVITIES

To further exploit the potential of EIS characterization, a specific Test Protocol for stack and system has been developed on the basis of intensive exchange with the system provider to allow the right comprehension of system specificities and operational constraints; during the exchange sessions limits for faulty conditions and the way to induce them have been discussed and agreed upon.

TEST CONCEPT SETUP

Advanced prognosis

FBK and SP focus on the development of the AI algorithms for forecasting the ROL of the co-generator’s water filter, one of the most critical components of the system. The main goal is the optimization of the maintenance procedure based on the specific usage of the co-generator, the external/environmental variables and derivatives from the signals. A dedicated WM has been deployed on the Microsoft cloud service Azure having all the software needed for a powerful and scalable AI framework. A first batch of algorithms has been developed for cleaning, arranging and manipulating the dataset to generate training/validation sets for the predictive algorithms. 

DATASET LIFESPAN

RUBY FOR SMART GRIDS and vpp

RUBY is developing advanced functions to implement energy management systems that would help achieving the optimal use of the energy within smart grids. UNISA is designing a supervisory strategy for an energy grid accounting for the three main energy vectors (electricity, heat, gas), final uses and production sites, with PEMFC and SOFC, including reversible technology. A multistate management algorithm will esxploit the MDPC tool (Monitoring, Diagnostic, Prognostic and Control) for  the optimization of the overal efficiency making use of virtual power plant (VPP) concept.

Picture1
Smart Grid configuration for the design of the optimal energy management via vpp
0
NEW PROTOTYPE
0
STACKS UNDER TESTING
0
SYSTEMS UNDER TESTING
0
STUDENTS INVOLVED

Pin It on Pinterest