Best FTC resources

I'm thinking about bying some books on Faut Tolerant Control (detection, diagnostics ( along with fuzzy, neural networks), etc) and was wondering which books are the best in these subjects. Also if someone has an input on any other resource I'll appreciate it.

Regards

Todd

Reply to
Todd Mercer
Loading thread data ...

I don't know if any one book will provide everything you mention. You might look at:

Control System Safety Evaluation and Reliability, 2nd Edition by W. M. Goble Practical Design of Safety-Critical Computer Systems by William R. Dunn Safety Shutdown Systems: Design, Analysis, and Justification by Paul Gruhn & Harry L. Cheddie

All are available from

formatting link

John Shaw

Reply to
John Shaw

Todd,

Fault tolerant control is actually standard practice in the process industries. I have used it for 25 years. Your original post gives the impression you have the preconceived idea that fuzzy logic and neural networks are somehow a solution to this question. There was a quote a few posts up the list, "Many are stubborn in pursuit of the path they have chosen, few in pursuit of the goal." - Friedrich Nietzsche. Neural nets act based on an accumulated history, i.e. experience. The occurrence of a fault in a control system is supposed to be a rare event. Thus NNs would not have any experience to go on. Fuzzy logic depends on a description of situations and responses just as hard logic does. The only difference is that some ambiguity is tolerated.

The fault tolerant systems I have dealt with had the following features:

1 - Fault tolerant sensors with diagnostics. This is the reason that process industries world wide have dumped process sensing switches in favour of transmitters.

2 - Multi-redundant sensors. A variety of selection schemes exist, e.g. 'middle of three', 'any two out of many', 'use A unless bad, then use B', '1oo2', '2oo2', '2oo3', and others.

3 - Dual redundant control electronics with diagnostics and full auto, bumpless transfer. This has been standard in DCSs for 25 years. Less successful with PLCs. (I don't know why.)

4 - Dual redundant communications with diagnostics and full auto, bumpless transfer. This has been standard in DCSs for 25 years. Less successful with PLCs.

4 - Dual redundant output channels such as twin solenoids.

5 - Diverse redundancy, especially on outputs.

Systems such as Honeywell FSC, Triconex TMR, Moore Quadlog are designed with such functionality however it can be user implemented on any platform. (Except item 3.)

The books listed below are written by designers of such systems.

Walter.

Reply to
Walter Driedger

Thank you John and Walter for your input and I read about those books on amabon and bn.com and seem very practical and will get them eventually. Any other suggestions on more theoretical books, let's say at the modeling level?

Regards,

Todd

Reply to
Todd Mercer

Todd,

You should be looking for works of Ron Patton. He has coauthored a few books on fault tolerant control and various methods of doing it. In the previous posts hardware redundancy approaches were illustrated. I would hazard a guess that those systems Walter described were actually fault operational, and not fault tolerant. Nevertheless, analytical redundancy is widely used, as well. Also, it is worth mentioning that fault tolerant control encompasses passive and active approaches. The former is kind of robust control, the latter makes use of active diagnostic information, which is performed by fault diagnosis methods.

Below are a few references for you to start. If you want some more information, please get in touch with me off the list.

Andrey

---------------------

Fault Diagnosis in Dynamic Systems: Theory and Application by Ron Patton, Paul Frank, Robert Clark (Editor) * Hardcover: 602 pages

  • Publisher: Prentice Hall; (November 1989)

Robust model-based fault diagnosis for dynamic systems / by Jie Chen, Ron J. Patton . - Boston : Kluwer Academic Publishers, cop. 1999. - XVIII, 356 p : il. - (The Kluwer international series on asian studies in computer and information science; 3) . - ISBN 0792384113

ISSUES OF FAULT DIAGNOSIS FOR DYNAMIC SYSTEMS Issues of fault diagnosis for dynamic systems / Eds. Ron J. Patton, Paul M. Frank, and Robert N. Clark . - London : Springer, cop. 2000. - XXV, 597 p : il . - ISBN 3540199683

GERTLER, Janos J Fault detection and diagnosis in engineering systems / Janos J. Gertler . - New York : Marcel Dekker, Inc, cop. 1998. - XX, 484 p : il .

- ISBN 0824794273

CHEN, Tinghuai Fault diagnosis and fault tolerance: a systematic approach to special topics / Tinghuai Chen . - Berlin : Springer-Verlag, cop. 1992. - XII,

197 p : il . - ISBN 3-540-54962-5
Reply to
Andrew Romanenko

A new phrase to me. Please explain the distinction. I do remember pulling boards out of my DCS controllers at random and watching the fail-over messages pop up but absolutely no process control interference. It sure seemed like fault tolerance to me.

Walter.

Reply to
Walter Driedger

Walter,

You are absolutely right that your system had fault tolerance. However, since you saw no changes when you caused a controller malfunction the setup was fault operational, that is, it could continue without any loss of performance. That is the best case and the most expensive one. Fault tolerant systems, in contrast, are able to maintain operation although with eventually degraded performance if a fault occurs. I think I won't be wrong if I say that most research activity in fault tolerant control concerns this type of systems, because they do not require redundant hardware but instead use analytical redundancy or the like. Finally, the third type of systems is fault-save, that is, those that are brought to a safe state in the event of a fault.

The above is based on the definitions of IFAC SAFEPROCESS commettee. I have just looked up some links and this one is very good:

formatting link
But according to this list your system may also be referred to as fault-tolerant. Well, the terminology dust hasn't settled yet :)

Andrey

Reply to
Andrew Romanenko

Todd Mercer wrote: "I'm thinking about bying some books on Faut Tolerant Control (detection, diagnostics ( along with fuzzy, neural networks), etc) and was wondering which books are the best in these subjects. Also if someone has an input on any other resource I'll appreciate it."

"Walter Driedger" responded: 'Your original post gives the impression you have the preconceived idea that fuzzy logic and neural networks are somehow a solution to this question. There was a quote a few posts up the list, "Many are stubborn in pursuit of the path they have chosen, few in pursuit of the goal." - Friedrich Nietzsche.'

My interpretation of the original post is "fuzzy logic and neural networks _may be_ a solution to this question".

"Walter Driedger" continues: "Neural nets act based on an accumulated history, i.e. experience. The occurrence of a fault in a control system is supposed to be a rare event. Thus NNs would not have any experience to go on."

This would be true only if the training data set is poorly designed. I don't see why rare events couldn't be handled through common statistical procedures, such as stratified sampling.

-Will Dwinnell

formatting link

Reply to
Will Dwinnell

Todd Mercer wrote: "I'm thinking about bying some books on Faut Tolerant Control (detection, diagnostics ( along with fuzzy, neural networks), etc) and was wondering which books are the best in these subjects. Also if someone has an input on any other resource I'll appreciate it."

Try the FAQs for comp.ai.neural-nets or comp.ai.fuzzy here on Usenet or Will's Technical Book List, at:

formatting link

-Will Dwinnell

formatting link

Reply to
Will Dwinnell

Todd Mercer wrote: "I'm thinking about bying some books on Faut Tolerant Control (detection, diagnostics ( along with fuzzy, neural networks), etc) and was wondering which books are the best in these subjects. Also if someone has an input on any other resource I'll appreciate it."

I forgot to mention in my other message that you might want to look into the sensor fusion literature. I'd suggest "Multi-Sensor Fusion: Fundamentals and Applications with Software" by Brooks and Iyengar (ISBN: 0139016538) as a start.

-Will Dwinnell, MBA

formatting link

Reply to
Will Dwinnell

Thanks for the link.

Walter.

Reply to
Walter Driedger

Thank you all for your useful suggestions!

TM

Reply to
Todd Mercer

PolyTech Forum website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.