Intelligent Supervision

 
image005.jpg

Modern society depends strongly upon the availability and correct operation of complex technological systems (cars, planes, trains, etc.) and processes (water and energy distribution networks, chemical factories, ...). Due to the simultaneously increasing economic demands and the numerous ecological and safety restrictions to be met, high dependability of those systems and processes has become a dominant goal in the recent years. However, malfunctions (or fault) can appear causing economic losses, damage to operators and machines, disturbance to the users, etc. In a general sense, a fault is any change in the behaviour of any component of the system.

In order to increase reliability, safety and availability of a system, the CER CS2AC-UPC works to develop new methodologies and intelligent supervision tools that make possible:

 
image002.jpg




image025.gif

Complex technological systems and processes are subject to a number of requirements:

  • They have to operate continuously over long periods of time and with a minimal intervention of the operator;
  • The cost derived from faults and problems maintenance must be reduced;
  • Must be supervised by operators that could be not very specialized and can be remotely controlled.

 

To fulfil these requirements a supervising system that monitors and improves the availability of the plant avoiding downtime and unplanned system failures was needed. The requirements of this system are:

  • They have to work in a continuous way for long periods of time, and with the minimum intervention of the operators;
  • The cost derived from failures and problems of maintenance has to be low;
  • It is necessary that they can be supervised by not really prepared and specialized operators and remotely controlled, since these plants are installed in a great diversity of countries.
 
On a universal level there are different collectives that work to give answer to this challenge. The CER CS2AC-UPC is part of some of these collectives that periodically meet to share and to discuss the advantages and inconveniences of the different solutions brought.  

More concretely, the research carried out by the CER CS2AC-UPC within this area is:
  • To develop representative models of the behavior of the systems or processes, capable of distinguishing between a normal behavior or a behavior in failure.
  • To define strategies based on models able to allow to locate failures.
  • To distinguish the degree of diagnosis of a system or process from the measurable variables.

 

 

IMG_5968.JPG
 
A way to increase the reliability and security of the automated processes consists on provinding them with mechanisms of tolerance. For this reason there is an increasing need and interest to develop systems of control, capable of operating in an acceptable and safe way, even after the appearance of a failure, and also able to stop the process before irreparable damages appear in it or in its scene. This kind of control systems are called Fault Tolerant Systems.

The tolerance to failures it is then understood as the capacity of a control system to maintain the objectives of control, in spite of the appearance of a failure, accepting a certain degradation of its benefits.

IMG_5962.JPG

In concrete, tolerant control (active) consists on having a system of diagnosis of failures that, in real-time, it would be able to inform about the appearance of a failure to a supervisory system, so that some correcting mechanism is activated. Concretely the system of diagnosis should determine the damaged component, the type of failure, its magnitude and the moment it occurred. From this information, the supervisory system should activate some mechanism of accommodation of it or of reconfiguration of the control, or even, depending on the importance of the failure,  the stopping of the system.
 
Control_tolerant_esquema.png

 
Another way of increasing the reliability, improving the availability of the machines and reducing the costs coming from the non desired stops, consist on implementing strategies of predictive maintenance, aiming to minimize the appearance of  failures, through the planning of the tasks of maintenance. This is achieved making the follow-up of the ageing of the system (incipient failures) and estimating the useful life of the components of the process. The supervisory computer systems that incorporate predictive maintenance associated to diagnosis tools, are called System Health Management (SHM).
 
The CER CS2AC-UPC is working in the development of tools which have integrated the diagnosis, the prediction and the control with the aim of achieving improvements in the availability, reliability, quality and security of the systems. A first development carried out is shown in the attached figure.

Esquema

 

These thecniques are being applied in a wide variety or real processes and laboratory, such as: wind turbines, UAVs, gas turbines, water distribution networks, fuel cell, ....