HARCRICS

Critical infrastructure systems (in short CIS), such as water, wind farms, gas or electrical networks, and complex systems, such as wind turbines, elevators or UAVs, are, in general, composed of a large number of elements with time-varying behaviour, exhibiting numerous operating modes and subject to changes due to external conditions (e.g., weather, load) and operational constraints. Currently, operational constraints of automated systems and process involve, in addition to the field of safety, availability and reliability, the field of health aware control. Health aware control deals with the ability to generate a control action to fulfil the control objectives and to extend the life of the system components.

The safe, reliable and health operation in this type of technological processes and complex system is of great significance, for the protection of human life and health, the environment, product quality and the invested economic value. In order to ensure that high performance can be archived and maintained under adverse conditions, they require highly sophisticated real-time supervisory and control systems. Due to the probability of failure of some of these components will increase exponentially with their number and complexity, operational designs must take into account mechanisms for early detection of faults avoiding performance degradation and damage to the machinery or human life. The goal of the resilient control system is to maximize the survivability of the critical infrastructures and systems when they are subject to the adverse operating conditions, such as component (hardware, software, or power supply) degradations and failures, natural disasters, human errors, malicious attacks, and environmental changes.

This project introduces new control paradigms, named “health-aware and resilient control” that considers the information about the system health to adapt the objectives of the control law to extend the remaining useful life, and in case a fault appears (due to natural disasters, accidental failures or malicious attacks) to activate resilient mechanisms to mitigate the fault effect. This new paradigm combines two concepts:

• the new concept of Health-Aware Control (HAC) that integrates the tasks of diagnostic  and prognostic modules, as well as processes and procedures responsible for information gathering about the system health enabling to make the right decisions for extending the system life in normal and in faulty situations, 
• and the modern concept of Resilient Control (RC) that tries to design control algorithms to allow the survivability in despite of faulty situations.

The new paradigm of HAC considers the information provided by the prognosis algorithms about the system health to adapt the controller in such a way that the control objectives/constraints are adapted according to the system health. In this way, the control actions are generated to fulfil the control objectives/constraints but at the same time to extend the life of the system components. So, HAC tries to achieve maximum performance while not degrading so much the system. The trade-off is based on solving a multi-objective optimization problem.   Controllers designed with standard procedures aim to maximize performance without caring about whether the resulting action will deteriorate very fast the system life.

 On the other hand, RC in a new concept based on the  idea introduced by (Rieger, 2009): “A resilient control system is one that maintains state awareness and an acceptable level of operational normalcy in response to disturbances, including threats of an unexpected and malicious nature”. 

To achieve this general goal several important issues will be addressed in the project development:

• Establish a general methodology for including in a control system health-aware and resilient mechanisms.
• Develop prognosis algorithms that can operate in real-time and can be embedded in the health-aware control by adapting the objectives/constraints of the controller.
• Develop health assessment algorithms that could be integrated with prognosis algorithms that can work in a collaborative way providing complementary information about the current and future health state of the system to alarm and support the predictive maintenance decisions based on the conditions estimation.
• Develop resilient mechanisms that can work in a complementary way with the health-aware control strategies that can guarantee the system operation in case of a faulty situation.
• Develop algorithms that can guarantee the quality of the sensor information used by the prognosis and diagnosis algorithms by means of data validation and reconstruction techniques.
• Develop algorithms for the optimal sensor/actuator placement that allow to deploy the health-aware and resilient control algorithms.
• Application of the developed algorithms in a challenging real-applications as critical infrastructures (water networks and wind farms), and complex systems (elevators and UAVs), that can serve as scientific and technological demonstrators for testing the algorithms and to show the technical and economical advantages of health-aware and resilient control.