This discussion is locked.
You cannot post a reply to this discussion. If you have a question start a new discussion

Introduce Yourself

Please help us to keep this MyCommunity space up to date and relevant by letting us know who you are, what you do and what you would like to see here?



Thanks

Joanne

  • Hello Joanne (and all),



    Introduction: I run a small control engineering consultancy (ISC Ltd) based in Glasgow, and we specialise in control design, modelling/simulation and optimisation. Recently we won Application of the Year for the Control System Software for a very large hydraulic mechanism used during the installation of wind turbines (see this: sine.ni.com/.../cs-15650 - an impressive mechanical design by Houlder). We work across many sectors - but always at the low level dynamic (feedback) control level whether for a power station or a car engine.



    I suppose one thing I find difficult with The IET C&A Network (I was on exec some years ago), is that it covers such a broad range of specialisms as the words "Control" and "Automation" can refer to almost anything (from the sort of stuff we do, to IT systems, to management / supply chains - all very different) and in almost every industry which have different needs in this area. For example, when IET C&A started doing a Water Industry event it was mostly on automation systems (the management of software, maintenance etc.) even though good low level control can be very important in that sector. Similarly most of things labelled "Control" I read about in the IET magazine, don't seem at all relevant to my interest. Maybe this is because the view is that this low level feedback control is a bit "old hat", but it remains vital to get it right for good operations and there are plenty of ways it can go wrong.



    But I do think its this broad nature of the subject group that is a bit of a problem - you can also see some of posts in this forum also seem quite disparate - and consequently we maybe go and look at forums that are more focussed and directly useful.



    So what I would like to do here is take part in vibrant discussions about control ...
  • Thanks Andrew

    You make some good points there and definitely food for thought for the next exec team meeting.


  • Perhaps the network should be called "IET Process Control and Automation Network"



    Revised IET Process Control and Automation Network Description:

    This community has been established to represent all those with an active interest in Process Control and Automation. It will be relevant to those involved with the design, implementation, construction, development, integration, analysis and understanding of process control systems and their operating environments.



    I don't think you can completely separate process control algorithm design from the complexity, reliability, testability and cost of the whole system within which they are designed to operate. This includes the design of the computational platform they run on. (Computational platforms of increasing complexity being for example - simple bespoke state-machine style firmware running on a dedicated micro-controller, simple proprietary PLC OS or full-blown preemptive multi-tasking operating system).



    As flexibility and complexity increases year on year, it becomes ever more difficult to design and build reliable, dependable, robust and safe process control systems. It seems that opportunity to decrease system development costs are being taken at the expense of long term robustness and reliability. Process control system complexity itself has to be controlled and partitioned in the right way to maintain and improve testability, maintainability and reliability.



    I think the complexity escalation problem has hit engineers like a tsunami in the last 20 years, especially with the advent of the Internet and the year on year increase in the flexibility, complexity and processing power of cheap computer hardware. As a result hackers, viruses and computer worms (such as Stuxnet) can seemingly get anywhere they please now in the digital world, to monitor, destroy or subvert ever larger classes of process control system. (The complexity tsunami is now reaching our electricity and gas meters, our voting systems etc., and as a community we need to get an intellectual grip of the problem before it washes us all away).



    I want to start talking about getting to grips with the complexity problem, especially in regards to designing reliable and safe process control systems for Generation IV nuclear reactors, especially molten salt nuclear reactors. 
  • Process Control is a big area with, as you say, some quite big shifts in complexity in recent year - both at the low level control tier right through to the the high level management information systems. But it is also has a very different set of language and usage/deployment to the Embedded Control found on cars, marine, aircraft etc. At the lowest level the dynamics can be described by common theories and good practice, but as soon as it is taken to the application the issues, priorities and activities around them are very different. There's also got a fairly strong distinction between continuous process control and batch process control.



    I 'm not sure how things are best defined within this TPN to cover all of these ... or even if all of these do need to be covered. It would be really good to see something that defines how these very different aspects of Control and Automation link together and which parts are receiving the attention needed by the wider community.
  • I am most familiar with Batch Control in terms of the manufacture of speciality chemicals or in the contol of inertia simulation dynamometers. Most batch control systems contain subsystems which require continuous control: these can be either true continuous real-time control processes or batch processes which start but never finish. In that philosophical sense batch control can be considered a superset that includes both batch and continuous control.



    Every industry seems to have its own terminology in terms of control (especially batch control), which doesn't help.
  • Former Community Member
    0 Former Community Member
    Hi Joanne,

    Looking at it another way round, maybe we could have discussion topics where people are not just posting their CV - there are other forums better suited for this. I switch off when I see this.



    I like James's first post on complexity and a discussion topic on methods of abstraction would be welcome.

    But you would need to mediate such topics to keep them on subject or it could be hijacked to meaningless discussion of continuous vs batch and people will disengage. Unless you can create sub-threads that people can choose to follow or abandon?



    If possible, maybe you could post a list of automation sub-topics and everyone could vote on them (need a web based voting mechanism I suspect) and we could kick off the ones people want to discuss.



    Cheers

    Alan
  • Abstraction is an interesting topic. Also sensor interfacing and simulation, both of which are helped by some forms of abstraction.



    When I attended a short course on neuroscience at the Royal Institution, I realised that the brain uses some of the abstraction techniques that software engineers commonly use to aid simulation and testing. I expect it uses lots more I didn't spot. I think the brain is very sophisticated in its ability to abstract, therefore if you want to talk about abstraction, it would be more than worthwhile having an event where control engineers talk with neuroscientists I think.



    In his difficult to read book "A Model of the Brain" (1964) J. Z.Young goes as far as trying to think about the "Brain as the Computer of a Homeostat". I think this is a reasonable approach to take to try and understand and talk about some of the lower level fucntions of the brain.



    http://en.wikipedia.org/wiki/Homeostat





    I'm not sure what elese people would find interesting to discuss. Serial control compared with parallel control perhaps?



    One of the interesting features of control and automation is that you often have to have a database attached - especially so for batch control. There are probably design elements to this union that could be improved and simplified. I did think on this many years ago but I have completely forgotten what conclusions I came to.
  • I've been following the discussion with interst, at one time I chaired the C&A TPN but am now retired and too far removed from the day to day developments to add much.



    However I was recently reading about the investment (£60m) being made in the Manufacturing Technology Centre in Coventry and wonder if the IET either through this TPN or the wider Design & Production Sector should develop a relationship with the MTC. From what I can see both industry and academia are involved but none of the Engineering Institutes.



    When the IET was formed this TPN looked at the idea of various interest/discussion streams but we did not have the IT system to make it possible and it was dropped. Now we have the tools it seems a good time to try it. I agree the discussions need mediating. As Alan said CV's and recruting agancies are not appropriate in this type of forum.



    I'll now let you poor soles go back to work, I'm of to the golf course :-)








  • I agree it was spug and as punishment I only managed to play 8 holes before the heavens opened :-(



    I know the time seems to have flown by. If I can help in anyway let me khow



    John
  • I would like to introduce myself as David Martin, but so far the IET's web team hasn't found a way of letting those of us not known by our first names label ourselves with the name everyone in the real world knows us by.