I am working on FTC for handling sensor faults in the system. I have read about observer-based FTC schemes which use error threshold-based logic between the states and estimated states. If a sensor is faulty then the error value crosses this threshold and it is an indication of sensor fault and then we take the state/output estimated data instead of actual sensor measurement.
However, my question is that, if we want to achieve sensor fault tolerance or if we have to use an observer, why can't we use the estimated state/output information always and why can't we eliminate the sensor permanently. Basically, I want to know what is the point in using observer-based FTC for sensor faults compared to the option of complete observer-based/ sensorless design?
Please give your specific suggestions in this regard.
Please, look at https://www.sciencedirect.com/science/article/abs/pii/S1270963819300021
The fact you can observe quatitirs do not mean you can control the system. Sensors help to quantify deviations from what is considered a standard behavior.
If you can organize your system so that you can identify a control diagram, then it will be easier to investigate the role of your subsystem within your system
Please, look at https://www.sciencedirect.com/science/article/abs/pii/S1270963819300021
The fact you can observe quatitirs do not mean you can control the system. Sensors help to quantify deviations from what is considered a standard behavior.
If you can organize your system so that you can identify a control diagram, then it will be easier to investigate the role of your subsystem within your system
- Badges
- Science topic
- Similar topics
- Methodology
- Sensors