DEOCS: Monitorización, diagnostico y control tolerante a fallos de sistemas ciberfísicos con métodos basados en datos
Nowadays, with the spread of new communication and computing devices, critical infrastructure systems such as water distribution networks, electric power systems or telecommunication networks are becoming cyber-physical systems (CPS), i.e., systems composed of physical entities controlled or monitored by computer-based algorithms.
The design, monitoring, control and security of CPS are becoming increasingly challenging as their size, complexity and interactions are steadily growing. CPS are complex large-scale systems that, in turn, require highly sophisticated supervisory-control strategies. Standard techniques often rely on system models with a limited use of the vast amounts of operating data available through the use of new ICT devices.
In DEOCS, the development of real-time control and diagnosis systems that not only use system models but also include the available historic and real-time data in the decision loops (data driven techniques) is proposed. State-of-the-art data analysis techniques, coming from statistics and machine learning fields, will be applied to real-time system monitoring, sensor data validation, fault diagnosis and fault tolerant control.
Goals of DEOCS are the development of:
• Pattern recognition techniques than can determine the actual operation mode of CPS and predict their behavior in the future.
• Data-driven methodologies for state estimation of cyber-physical systems.
• Hybrid architectures that exploit the complementarities between data-driven and model-based fault diagnosis methods.
• Distributed data-driven strategies that allow monitoring and diagnosing large-scale CPS.
• Strategies for including fault tolerance in data-driven control techniques exploiting the spatial and temporal redundancy available in the system
• Data validation methodologies to guarantee the reliability of the proposed data driven monitoring and control methods.
• Methodology and performance indicators for assessing the impact of the proposed monitoring and control solutions.
The final goal of DEOCS is the validation of the developed techniques though their application to two real case studies: the Besòs waste treatment plant and the Barcelona’s drinking water network.
DEOCS is a project of the Spanish Ministry of Economy, Industry and Competitiveness, funded by Spanish State Research Agency (AEI) and the European Regional Development Fund (ERFD).
J. Barreiro-Gomez, C. Ocampo-Martínez and N. Quijano. On the communication discussion of two distributed population-game approaches for optimization purposes, 20th IFAC World Congress, 2017, Toulouse, France, IFAC, to appear.
J. Barreiro-Gomez, C. Ocampo-Martínez and N. Quijano. Partitioning for large-scale systems: A sequential distributed MPC design, 20th IFAC World Congress, 2017, Toulouse, France, IFAC, to appear.
M. Pourasghar, V. Puig, C. Ocampo-Martínez and Q. Zhang. Reduced-order Interval-observer design for dynamic systems with time-invariant uncertainty, 20th IFAC World Congress, 2017, Toulouse, France, IFAC, to appear.